Journal of Immunological Methods 238 (2000) 81–85
`
`www.elsevier.nl/locate/jim
`
`Radioiodination of human interferon-a2 interferes with binding
`of C-terminal specific antibodies
`
`a
`b
`a
`a
`´
`´
`Hana Schmeisser , Eva Kontsekova , Iveta Vancova , Vojtech Mucha ,
`a ,*
`Peter Kontsek
`aInstitute of Virology, Slovak Academy of Sciences,84246 Bratislava, Slovakia
`bDepartment of Microbiology and Virology, Comenius University,84215 Bratislava, Slovakia
`
`Received 5 July 1999; received in revised form 19 October 1999; accepted 5 January 2000
`
`Abstract
`
`I-interferon (IFN)-a2 are
`Radioimmunoassays based on reactivity between a monoclonal antibody (mAb) and human
`frequently exploited in interferon research. In general, epitopes of antibodies specific for human IFN-a2 are located on the
`two immunodominant structures formed in the N- and C-terminal domains, respectively. We found that labelling of IFN-a2
`125
`with Na
`I by the chloramine-T method did not affect the binding of antibodies recognising the N-terminal region 30–53. In
`contrast, radioiodination of IFN was associated with a dramatic decrease in IFN reactivity with mAbs specific for the
`C-terminus (residues |120–145|). We suggest that steric hindrance araising from the incorporation of
`125
`I into the tyrosine
`residues at positions 123, 130 and 136 may be responsible for the change in immunoreactivity. The adverse effect of
`radioiodination of IFN-a2 on the binding potency of C-terminal specific mAbs must be taken into consideration in
`experiments based on the interaction of such antibodies (i.e. NK2) with the labelled antigen.
`2000 Elsevier Science B.V.
`All rights reserved.
`
`125
`
`Keywords: Antigenic structure; Interferon-a2; Monoclonal antibody; Radioimmunoassay; Radiolabelling
`
`1. Introduction
`
`Human interferon (IFN)-a2 is the most frequently
`studied and therapeutically exploited type I IFN.
`Radioiodination of
`IFN-a2 by the chloramine-T
`method is a widely used technique for both structural
`and functional characterisation of this cytokine, i.e.
`mapping of antigenic sites on the IFN-molecule,
`screening for antibodies induced post-therapy and the
`analysis of interactions between the cytokine and its
`
`*Corresponding author. Fax: 1421-7-5477-4284.
`E-mail address: virupeko@nic.savba.sk (P. Kontsek)
`
`receptor. In iodine labelling, the iodine atom is not
`part of the structure of the natural molecule and
`therefore there are likely to be some differences in
`behaviour between radioiodinated and unlabelled
`materials (Bolton, 1977).
`There are two immunodominant structures iden-
`tified on IFN-a2, which are recognised by most of
`the specific monoclonal antibodies (mAbs). One
`structure is located in the N-terminal domain cen-
`tered around amino acids |30–40| and the other is
`located in the C-terminal domain spanned by resi-
`dues |112–148| (Kontsek et al., 1991). Comparing
`the reactivity of mAbs with IFN-a2 in a radioim-
`
`0022-1759/00/$ – see front matter
`PII: S0022-1759( 00 )00155-1
`
`2000 Elsevier Science B.V. All rights reserved.
`
`Page 1 of 5
`
`YEDA EXHIBIT NO. 2050
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`(cid:211)
`(cid:211)
`

`
`82
`
`H. Schmeisser et al. / Journal of Immunological Methods 238(2000)81–85
`
`munoassay (RIA) and in an enzyme linked immuno-
`sorbent assay (ELISA) we have observed distinct
`binding pattern for N-terminal and C-terminal spe-
`cific antibodies. We found that radioiodination of
`IFN-a2 was associated with a sharp drop in the
`binding capacity of antibodies directed to the C-
`terminal domain.
`In the present study we have
`addressed this topic in more detail since we believe
`that our results could be useful for the design and
`evaluation of experiments based on the interactions
`of C-terminal specific mAbs (the commonly used
`125
`mAb NK2 also recognises this region) with
`I-IFN-
`a2.
`
`2. Materials and methods
`
`2.1. Interferon
`
`Recombinant Escherichia coli-derived human
`8
`IFN-a2c (specific activity 10 U/mg) was kindly
`provided by Dr. G.R. Adolf (Bender, Vienna, Aus-
`tria).
`
`2.2. Monoclonal antibodies against human IFN-a2
`
`The preparation and characterisation of murine
`mAbs against
`IFN-a2 was described previously
`´
`(Kontsek et al., 1991, 1993; Kontsekova et al., 1992;
`Karayianni-Vasconcelos et al., 1993). Four N-termi-
`nal specific antibodies: N7 (epitope 36–38), N40,
`7/1 (epitope 30–41), N27 (epitope 43–53), and four
`C-terminal specific mAbs: N22, N39, N54 (epitope
`112–148), 2–19 (epitope 93–166) were used as
`hybridoma culture supernatants. MAb G23 specific
`for human IFN-g was used as a negative control. All
`antibodies were of
`the IgG1 isotype. The mAb
`concentrations were determined by ELISA using
`affinity purified mouse IgG1 as a standard (Kont-
`´
`sekova et al., 1992).
`
`2.3. RIA
`
`I by
`Recombinant IFN-a2 was labelled with Na
`the chloramine T method without loss of antiviral
`activity (Kontsek et al., 1991). Immunoplates (Maxi-
`Sorp, Nunc) were coated overnight at 378C with
`sheep anti-mouse F(ab9) -fragment (Sigma, 1 mg /
`2
`
`125
`
`ml, 50 ml/well in phosphate buffered saline — PBS,
`pH 7.2) and then saturated with 1% non-fat dried
`milk. Hybridoma supernatants were added (50 ml /
`well) and the plates were incubated for 1 h at 378C.
`5
`After washing, 50 ml of radiolabelled IFN (0.5310
`cpm) were added to the wells for 1 h atroom
`temperature. Following washing the radioactivity
`bound to each well was measured in a gamma
`counter. All antibodies were assayed in triplicate.
`Between each step the plates were washed with PBS
`containing 0.05% Tween 20.
`
`2.4. Determination of mAb-affinity by RIA
`
`The binding affinities of mAbs to IFN-a2 were
`determined in a competitive solid-phase RIA. For
`this purpose,
`the most simple model of antigen–
`antibody interaction, that based on the law of mass
`action, was chosen (Mucha, 1993). Briefly, increas-
`ing concentrations of the unlabelled antigen (13
`212
`210
`10
`–5310 M) in 50 ml PBS were mixed with
`5
`125
`50 ml PBS containing 10 cpm of
`I-IFN-a2. The
`mixtures were transferred on mAb-immobilised
`plates for 1 h at room temperature. After washing,
`the bound radioactivity was counted, competition
`curves for each mAb were plotted and the respective
`´
`affinities for IFN-a2 were calculated (Kontsekova et
`al., 1992). The results obtained represent average
`values from samples at
`three competitor concen-
`trations near the half inhibition point.
`
`2.5. ELISA
`
`(MaxiSorp, Nunc) were coated
`Immunoplates
`overnight at 378C with IFN-a2 (1 mg/ml, 50 ml/ well
`in PBS pH 7.2). Plates were saturated for 1 h with
`1% non-fat dry milk in PBS. The hybridoma super-
`natants were added (50 ml/well) and the mixtures
`incubated for 1 h at 378C. Bound antibodies were
`detected after
`incubation for 1 h at 378C with
`peroxidase-conjugated swine antibodies specific for
`mouse Ig (USOL, Czechia) diluted 1/1000 in PBS.
`Positive reactions were visualised by the addition of
`orthophenylenediamine and absorbance values were
`measured at 492 nm (A
`). All mAbs were tested in
`492
`triplicate. Between each step, the wells were washed
`with PBS containing 0.05% Tween 20.
`
`Page 2 of 5
`
`YEDA EXHIBIT NO. 2050
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`H. Schmeisser et al. / Journal of Immunological Methods 238(2000)81–85
`
`83
`
`2.6. Relative mAb-affinities determined by ELISA
`
`Immunoplates were coated overnight with two
`fold dilutions of IFN-a2 in PBS (starting concen-
`tration 1 mg/ml, 50 ml/well) and saturated with 1%
`non-fat dry milk. The concentrations of mAbs in
`culture supernatants were balanced to 2 mg/ml and
`then were added (50 ml/well) for 1 h at 378C. All
`subsequent steps were the same as described above
`for ELISA. Two independent measurements were
`performed.
`The rate of change of the absorbance vs. log[IFN-
`a2 dilution] was plotted for each mAb. The inflec-
`tion points of plotted curves were found to be
`localised around the absorbance value A 50.4.
`492
`Therefore, a titre of a mAb was estimated as the
`reciprocal dilution of IFN-a2 giving the absorbance
`A 50.4. Since the concentrations of the mAbs
`492
`were balanced, ranked titres refer to the relative
`affinities of respective antibodies.
`
`3. Results and discussion
`
`The majority of mAbs generated against human
`IFN-a2 can be classified into two groups, namely
`those specific for N-terminal epitopes and those
`specific for C-terminal epitopes. During experiments
`with these antibodies we have observed that both
`
`groups differ in their patterns of reactivity in the
`ELISA and RIA. To elucidate this phenomenon we
`performed study with eight mAbs against IFN-a2.
`Four antibodies recognised different epitopes located
`in the N-terminal domain (residues 30–53) whereas
`the epitopes of four others were located in the
`C-terminal domain (residues |112–148|). In the
`enzyme immunoassay, the N-terminal and C-termi-
`nal specific mAbs exhibited a similar reactivity with
`the solid-phase bound antigen (Fig. 1). However, in
`the radioimmunoassay a different pattern of reactivi-
`ty was observed. Compared to the N-terminal spe-
`cific mAbs, the potency of C-terminal specific anti-
`bodies to bind radiolabelled IFN was decreased
`about
`ten fold. To characterise this effect more
`precisely we performed experiments comparing the
`affinities of these mAbs as measured by RIA and
`ELISA.
`Firstly, mAbs to IFN-a2 were ranked according to
`their affinities as estimated by competitive RIA using
`radiolabelled IFN (Fig. 2). Only the affinities of
`N-terminal specific antibodies could be determined
`because the low binding capacity of C-terminal
`specific mAbs did not permitt the calculation of their
`affinities. Subsequently, an ELISA with immobilised
`IFN-a2 was used to rank the affinities of both
`N-terminal and C-terminal specific antibodies. In
`contrast to the RIA experiments, the data obtained
`did not indicate any difference in the affinities of the
`
`Fig. 1. Reactivity of N-terminal and C-terminal specific mAbs with human IFN-a2 in ELISA and RIA procedures. Data from one
`representative experiment. Error bars indicate standard deviation of the measurements.
`
`Page 3 of 5
`
`YEDA EXHIBIT NO. 2050
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`84
`
`H. Schmeisser et al. / Journal of Immunological Methods 238(2000)81–85
`
`Fig. 2. Ranking of affinities for C-terminal and N-terminal
`specific mAbs estimated by ELISA and RIA.
`
`two mAb-groups. Ranking comparisons showed that
`the affinities of N-terminal specific mAbs determined
`either by ELISA or RIA were comparable. On the
`other hand, when comparing the corresponding
`ELISA- and RIA-estimated affinities for C-terminal
`specific mAbs, a dramatic drop in their affinities for
`125I-IFN-a2 was observed. These data suggest some
`steric hindrance by the labelled antigen preventing
`the interaction of the C-terminal specific antibodies
`with their respective epitopes. What could be the
`nature of such interference? The chloramine-T meth-
`od is the most widely used method for the radioiodi-
`125
`nation of small amounts of protein. Na
`I is oxi-
`dised by chloramine-T in the presence of the protein
`to be labelled, with the subsequent incorporation of
`125I into the tyrosine residues of the protein (Bolton,
`1977). All five tyrosines of IFN-a2 can be localised
`in the three-dimensional model of this cytokine (Fig.
`3). Three are located at positions 123, 130 and 136
`in the D helix which represents a part of the C-
`terminal immunodominant region. Moreover, in the
`chloramine-T reaction tyrosine residues on the (im-
`mune) surface of the protein molecule iodinate most
`readily (Dube et al., 1964). In contrast,
`the N-
`terminal domain of IFN lacks such residues and the
`two remaining tyrosines at positions 86 and 90 are
`located in the C helix, which exhibits low immuno-
`genicity (Kontsek, 1994). Taking into consideration
`both theoretical and experimental data we suggest
`125
`that substitution of
`I for hydrogen atoms in the
`tyrosine residues of IFN-a2 leads to steric hindrance
`for mAbs with epitopes located in the C-terminal
`domain. However, we do not expect significant
`
`Fig. 3. Conformational model of human IFN-a2 constructed from
`the co-ordinates from the Brookhaven Protein Data Bank, Acces-
`sion Number 1RH29 (Radhakrishnan et al., 1996). N-terminal
`(30–53) and C-terminal (120–145) immunodominant structures
`are highlighted. The positions of five tyrosines are indicated (open
`circles).
`
`125
`
`conformational modifications of the domain because
`radiolabelling is not associated with any decrease in
`the biologic activity of IFN-a2 (Kontsek et al.,
`1991).
`I resulted
`Labelling of human IFN-a2 with Na
`in a sharp decrease of the reactivity of iodinated
`antigen with the C-terminal specific mAbs. This fact
`must be taken into consideration when selecting
`125
`mAbs for experiments with
`I-IFN-a2. In other
`type I IFNs (even of different species) the hydro-
`philic character of
`the D helix suggests strong
`immunogenicity and the potential for eliciting mAbs.
`Because of the number of tyrosine residues in this
`helix (e.g. human IFN-b has four),
`it
`is to be
`expected that
`iodination of mAbs specific for D-
`
`Page 4 of 5
`
`YEDA EXHIBIT NO. 2050
`MYLAN PHARM. v YEDA
`IPR2015-00644
`
`

`
`H. Schmeisser et al. / Journal of Immunological Methods 238(2000)81–85
`
`85
`
`helices will be deleterious. Therefore when selecting
`mAbs for binding to IFN (or cytokines in general)
`following chloramine-T radiolabelling, their epitopes
`should be examined for the proximity of tyrosine
`residues.
`
`Acknowledgements
`
`This work was supported in part by the Slovak
`Grant Agency for Science VEGA (Grant No. 2 /
`5039/98, 2/5041/99 and 1/7377/ 20).
`
`References
`
`Bolton, A.E., 1977. Radioiodination techniques. In: The Radio-
`chemical Centre, Amersham, Bucks, UK, Review 18.
`Dube, S.K., Roholt, O.A., Pressman, D., 1964. Identification of the
`most rapidly iodinating tyrosine residue in a-chymotrypsin. J.
`Biol. Chem. 239, 3347.
`
`ˇ
`´
`Karayianni-Vasconcelos, G., Fuchsberger, N., Skamlova, Z., Kont-
`sek, P., 1993. Comparison of antigenic properties of three
`interferon (IFN)-a2 subvariants and establishment of a quan-
`titative IFN-a2 ELISA. Acta Virol. 37, 509.
`Kontsek, P., 1994. Human Type I
`interferons: structure and
`function. Acta Virol. 38, 345.
`´
`´
`´
`´
`Kontsek, P., Borecky, L., Kontsekova, E., Macikova, I., Kol-
`´
`´
`´
`cunova, A., Novak, M., Krchnak, V., 1991. Mapping of two
`immunodominant structures on human interferon a2c and their
`role in binding to cells. Mol. Immunol. 28, 1289.
`´
`Kontsek, P., Borecky, L., Zavyalov, V.P., Maiorov, V.A., 1993.
`Peptide-mapping of three neutralising epitopes into predicted
`biologically active sites of human interferon-a2. Immunol.
`Lett. 35, 281.
`´
`´
`Kontsekova, E., Kolcunova, A., Kontsek, P., 1992. Quadroma-
`secreted bi(interferon-a2-peroxidase) specific antibody suit-
`able for one-step immunoassay. Hybridoma 11, 461.
`Mucha, V., 1993. An operational model of the antigen–antibody
`interaction. Acta Virol. 37, 388.
`Radhakrishnan, R., Walter, L.J., Hruza, A., Reichert, P., Trotta,
`P.P., Nagabhushnan, T.L., Walter, M.R., 1996. Zinc mediated
`dimer of human interferon-a2b revealed by X-ray crystal-
`lography. Structure 4, 1453.
`
`Page 5 of 5
`
`YEDA EXHIBIT NO. 2050
`MYLAN PHARM. v YEDA
`IPR2015-00644