- Email: [email protected]
Multiplexing analysis of the polyspecific intrathecal immune response in multiple sclerosis
Methods 56 (2012) 528–531
Contents lists available at SciVerse ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
Multiplexing analysis of the polyspecific intrathecal immune response in multiple sclerosis Alina Kułakowska a, Barbara Mroczko b, Maria Mantur c, Natalia Lelental d, Joanna Tarasiuk a, Katarzyna Kapica-Topczewska a, Ute Schulz d, Peter Lange e, Rüdiger Zimmermann d, Johannes Kornhuber d, Piotr Lewczuk d,⇑ a
Department of Neurology, Medical University of Białystok, Poland Department of Biochemical Diagnostics, Medical University of Białystok, Poland Department of Clinical Laboratory Diagnostics, Medical University of Białystok, Poland d Department of Psychiatry and Psychotherapy, Universitätsklinikum Erlangen, Germany e Lab for Clinical Neurochemistry, Department of Neurology, University of Göttingen, Germany b c
a r t i c l e
i n f o
Article history: Available online 16 March 2012 Keywords: Multiple sclerosis Clinical neurochemistry Cerebrospinal fluid MRZH-reaction Multiplexing Neuroimmunology
a b s t r a c t Intrathecal synthesis of the antibodies specific to neurotrofic viruses: measles (M), rubella (R), VaricellaZoster (Z), and/or H. simplex (H), known as ‘‘MRZH-reaction’’ plays important diagnostic role in multiple sclerosis (MS). Whereas the analysis of the oligoclonal IgG bands provides high sensitivity, the MRZHreaction shows high specificity, and hence these methods complement each other. For the first time we applied multiplexing bead-based technology to simultaneously analyze cerebrospinal fluid (CSF) and serum concentrations of antibodies against these viruses, and to calculate the antibody specific indices (ASI’s). The method shows reasonable precision: intra-assay, 2.9–6.7%, and inter-assay, 2.0–3.2%. The results are comparable with these obtained with other methods (ELISAs), including two runs of the certified external quality control schemes. Eighty-one percent of the MS cases (n = 27) and none of the sexand age-matched controls (n = 14), except one subject with ‘‘borderline’’ anti-measles ASI of 1.5, showed intrathecal synthesis of IgG against at least one of the viruses discussed. The ratios of the MRZH-positive cases in the MS group were: 12/22 for M, 12/19 for R, 13/26 for Z, and 7/26 for H. We conclude that the multiplexing technology can be applied as a tool to study the intrathecal immune response in the diagnosis of MS. Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction Twenty years ago, the intrathecal polyspecific immune reaction was described for the first time as a specific diagnostic tool in multiple sclerosis (MS) [1]. The detection of the intrathecal IgG against: measles (M), rubella (R), Varizella-Zoster (Z), and/or H. simplex (H) viruses in MS turned out to be of different quality than the detection of the oligoclonal IgG bands (OCB) on the isoelectrofocusing (IEF): whereas the presence of OCB is very sensitive, and MS subjects without them are extremely rare, it is not specific at all. On the other hand, detection of the intrathecal synthesis of the M, R,
Abbreviations: ASI, antibody specific index; CSF, cerebrospinal fluid; H, Herpes simplex virus; IEF, isoelectric focusing; M, measles virus; MS, multiple sclerosis; OCB, oligoclonal IgG bands; R, rubella virus; Z, Varicella-Zoster virus. ⇑ Corresponding author. Address: Lab for Clinical Neurochemistry and Neurochemical Dementia Diagnostics, Department of Psychiatry and Psychotherapy, Schwabachanlage 6, 91054 Erlangen, Germany. Fax: +49 9131 85 34238. E-mail address: [email protected] (P. Lewczuk). 1046-2023/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymeth.2012.03.002
Z, and/or H viruses, correspondingly frequently referred to as ‘‘MRZ-’’ or ‘‘MRZH-reaction’’, provides specific information already at the moment of the first clinical symptoms [2,3]. The growing number of potentially important biomarkers to be analyzed in a limited volume of the cerebrospinal fluid (CSF) makes techniques of a simultaneous analysis of several parameters in a single small-volume CSF sample a method of choice in the future [4]. Such technologies are generally defined as ‘‘multiplexing’’ independently of the physical–chemical background of the method. For example, the flow cytometric-based Luminex xMAP technology involves coupling of specific capturing ligands to the surface of microsphere (bead) sets uniquely identified with a combination of two fluorescence dyes [5,6]. This allows simultaneous reaction with up to one hundred (theoretical limit) antigens in a single sample. Recently, we [7,8] and others [9,10] applied Luminex-based technology to simultaneously measure several biomarkers of neurodegeneration and neuroinflammation. Following these studies, in this report we applied for the first time a multiplexing technology to simultaneously analyze the intrathecal generation of the
A. Kułakowska et al. / Methods 56 (2012) 528–531
antibodies against M, R, Z, and H viruses in patients with multiple sclerosis. 2. Materials and methods 2.1. The assay and the Luminex platform setting The concentrations of the specific antibodies were measured with the Serion Multianalyt™ bead-based assay (Institut Virion/ Serion GmbH, Würzburg, Germany), which enables simultaneous measurement of the concentrations of IgG specific against measles (M), rubella (R), Varizella-Zoster (VZV), and Herpes simplex (H) viruses in a small volume of a serum or a CSF sample. Serum (diluted 1:100 if not stated otherwise) and CSF (diluted 1:2 if not stated otherwise) samples from a given patient were analyzed (in duplicate) always on the same plate and in the same analytical run. To wash the filter plates, an automated vacuum washer (ELx50, BioTek, Bad Friedrichshall, Germany) was used. Bead populations were sorted, and the signal intensities were measured with the Luminex reader (Austin, USA). Serion Multianalyt™ assays use microspheres (beads) with fluorescence color-coding different than other Luminex-assays. Therefore, the Luminex 100 IS software had to be adjusted to enable raw data acquisition and storage. These data, stored as ‘‘run-files’’, were then analyzed by the Serion Multianalyt™ easyPLEX software. The Luminex 100 IS software was set to analyze at least 100 beads/population with the sample volume set to 60 lL and the sample timeout set to 40 s. 2.2. Linearity: intra- and inter-assay imprecision and external quality control (EQC) To check for the linearity of the measurements, a pooled serum sample was serially diluted with the assay buffer from 1:10 up to 1:1280 with the step of 2 (eight dilution points). To analyze intra-assay imprecision, two human CSF samples (diluted 1:2, and 1:8) and two serum samples (diluted 1:100 and 1:200) were assayed on one plate eight to ten times. The results are presented as the corresponding coefficients of variation (CV’s) of the concentrations of the antigen-specific IgG’s. Inter-assay imprecision was checked by the analysis of the median fluorescence intensity (MFI) of the signal of the positive control sample, included in the assay, measured in six repetitions on six different days. The results are presented as the corresponding CV’s. External (inter-laboratory) quality control was performed by the comparison of the results obtained in our laboratory with the consensus values obtained by n = 16–26 laboratories participating in two runs of the EQC scheme provided by the ISO15189-accredited Society for Promotion of Quality Assurance in Medical Laboratories (Instand, Düsseldorf, Germany). Since our laboratory was the only one participant using the multiplexing platform, our results were compared to, and had to agree with, the consensus values obtained on the ELISA platform. The results of the EQC are presented as the corresponding percentage deviations of our results from these consensus values.
529
2.4. Patients and the routine CSF/serum analyses All the CSF/serum samples were collected at the Department of Neurology, Medical University of Białystok, following the approval of the local ethical committee. All patients gave informed written consent for the study, and all of them underwent the lumbar punctures (LP) for the routine diagnostic procedures. Analyzed were CSF/serum sample pairs from 27 MS (average age, 39.8 ± 13.7 year, 21 females) and 14 sex- and age-matched control patients (average age, 37.6 ± 14.0 year, 11 females). The degree of the neurological impairment in the MS patients was evaluated using the Expanded Disability Status Scale (EDSS, [11]). According to the McDonald criteria [12,13], 21 patients had the relapsing–remitting form of the disease (RR-MS), and six patients had the primary progressive form (PP-MS). None of the MS patients was treated with any diseasemodifying drugs or corticoids at the time of the LP. MRI was performed in all MS subjects. Control group (n = 14) included patients with excluded neuroinflammation, namely: idiopathic headache (n = 6), idiopathic (Bell’s) facial nerve palsy (n = 3), ischialgia due to discopathy (n = 3), and polineuropathy (n = 2). After the LP, the CSF and blood samples were centrifuged (2000g, 20 min), and the supernatants of the CSF and the serum samples were aliquoted, frozen at 80 °C, and shipped to Erlangen on dry ice. IEF was performed with the Hydragel 9 CSF system (Sebia, Fulda, Germany) on agarose gel to fractionate the proteins, followed by immunofixation with peroxidase labelled anti-IgG to detect the OCB. Albumin and IgG concentrations were measured with immune-based nephelometry (Immage, Beckman Coulter) with the CSF and serum samples analyzed always in the same analytical run and correlated to the same calibrator. All MS showed intrathecal synthesis of IgG either on the IEF (the OCB pattern of the type ‘‘2’’ or ‘‘3’’ according to the consensus paper, [14]) or in the quantitative analysis according to Reiber [15]. Control patients showed neither clinical nor neurochemical symptoms of neuroinflammatory disorder; in particular, intrathecal IgG synthesis was not detected in these subjects. 2.5. Calculations of the antibody specific indices (ASI’s) If not stated otherwise, the data are expressed as the ASI’s, calculated according to Reiber [1,16]. Briefly, the concentration of the antibody of a given class (IgG in this study) against a given antigen (M, R, VZV, or H, respectively) in the CSF sample was divided by the concentration of the specific antibody in the corresponding serum sample (after adjusting with the dilution factors). So obtained ‘‘specific IgG quotient’’ was then divided either by the total IgG concentration quotient (QIgG), if no intrathecal synthesis was observed on the corresponding IgG-Reibergram, or by the upper limit of the total IgG concentration quotient (QIgG-Lim), if the intrathecal IgG synthesis was observed on the Reibergram (i.e. when QIgG > QIgG-Lim). ASI’s within the range of 0.6–1.5 were considered ‘‘normal’’, i.e. showing no evidence for the intrathecal IgG synthesis against given virus. The results higher than 1.5 were considered as suggesting the intrathecal specific IgG synthesis against a given antigen, and the results lower than 0.6 were considered implausible.
2.3. Validation of the assay: comparison with the ‘‘classic’’ ELISAs
3. Results
Five CSF/serum sample pairs were assayed with multiplexing in our laboratory, and their fresh aliquots were shipped to the Laboratory for Clinical Neurochemistry, University of Göttingen, where the analyses were re-performed with ELISAs (Dade Behring, Marburg, Germany) according to the routine protocols, and without prior knowledge of the multiplexing results.
3.1. Assay performance: linearity, imprecision, and the EQC results The results of the serum sample dilution experiment are presented in the Fig. 1. The curves show linear performance of the assay at least in the range of the dilution 1:40–1:640. The dilutions 1:10 and 1:20 resulted in the signals above the highest standard
530
A. Kułakowska et al. / Methods 56 (2012) 528–531
was normal (1.0), and in the S5 sample the pathologic H-ASI in the ELISA met the normal value of the multiplexing. In both cases, however, the overall interpretation of the results remained the same independently of the method, due to the agreement of the remaining ASI’s. 3.3. Polyspecific intrathecal immune response in patients; plausibility of the results The results of the MS and control patients are presented in the Table 2. Eighty-one percent of the MS cases and none of the controls (with one ‘‘borderline’’ M-ASI of 1.5 in one subject) showed the intrathecal synthesis of IgG against at least one antigen. In one MS and in one control patient, the ASI were lower than 0.6 (M, and VZV, respectively), otherwise all the results were plausible.
Fig. 1. The dilution curves prepared by a serial dilution of a pooled serum sample. The arbitrary concentrations in the sample diluted 1:10 is defined as ‘‘1000’’. Each point represents the average of double measurements. Dashed vertical line indicates the dilution 1:100, i.e. the usual dilution of a serum sample. MFI, mean fluorescence intensity.
4. Discussion point, whereas the dilution of the serum sample 1:1280 resulted in still measurable signals (i.e. distinguishable from the blank). Correspondingly, the typical dilution, as suggested by the manufacturer and applied in most of the samples of this study (1:100, indicated with a dashed line on the Fig. 1), is in the middle of the linear part of the curve. The CV’s of the intra-assay imprecision were: 4.4–6.7% for measles, 4.0–4.7% for rubella, 2.9–6.4% for VZV, and 4.4–5.9% for H. simplex, with better performance when the CSF sample was diluted 1:2 compared to 1:8, and comparable results when the serum sample was diluted 1:100 or 1:200. The inter-assay CV’s of the MFI of the signals of the control sample assayed on six different plates were: 2.5% for measles, 2.0% for rubella, 2.6% for VZV, and 3.2% for H. simplex. In two runs of the EQC scheme, the deviations of the results obtained in our laboratory from the consensus values were: +8% and 17% for measles, +4% and 17% for rubella, 20% and 16% for VZV, and +4% and 10% for H. simplex. In both runs all the results were found acceptable by the EQC scheme provider, and all the diagnose-oriented interpretations based on the results obtained were correct.
In this report, for the first time we present the application of the multiplexing technology to simultaneously quantify humoral immune reaction against neurotrofic viruses (measles, rubella, Varizella-Zoster, and H. simplex) as a diagnostic tool in multiple sclerosis. This phenomenon, originally described about 20 years ago by Reiber [1] has been since then referred to as ‘‘polyspecific intrathecal immune response’’ or ‘‘MRZ(H) reaction’’. Compared to the original results reported by Reiber [1], we observed almost identical ratios of the MS patients with the intrathecal synthesis of the IgG against: rubella (63% vs. 60%), VZV (50% vs. 55%), and H. simplex (27% vs. 28%), with somehow lower ratio of the MS patients positive for measles (55% vs. 78%). This discrepancy, which resulted in a lower overall ratio of ‘‘positive’’ subjects in our study (81% vs. 90%), might be explained for example by the geographical and generational differences of the populations studied (our MS patients were born about 20 years after these studied by Reiber), resulting in different anti-measles systemic immune responses, for example due to different anti-measles immunization schemes. Simultaneous analysis of many biomarkers with multiplexing technology has several advantages compared to the classic oneparameter methods (like ELISA): (a) it saves precious sample volume, which is particularly important if body fluid difficult to obtain are analyzed (like the CSF); (b) it reduces workload (one run instead of four, for the MRZH-reaction, or even dozens for other applications); (c) it improves the quality control, as it is easier to confirm or to exclude mistakes, like pipetting errors; (d) it improves plausibility control and the diagnosis-oriented interpretation of the results. Particularly the latter point is important for the MRZH reaction: in theory, lack of the intrathecal immunoglobulins synthesis should result, ex definitio, in all ASI’s equal to 1.0, or less conservatively spoken, all the ASI’s in a given CSF/serum sample pair should be more or less the same. This, however, requires the assumption that all the corresponding reactions are performed under precisely the same conditions (the same temperature and
3.2. Comparison of the multiplexing results with the ELISAs The between-methods comparisons of the results are presented in the Table 1. In four out of five sample pairs, not only similar ASI’s were obtained with both methods performed in two different laboratories, but perhaps even more importantly, the diagnose-oriented interpretation (the presence or the absence of the polyspecific intrathecal immune response) was identical: twice ‘‘negative’’ and twice ‘‘positive’’. In one sample (S2), the interpretation of the multiplexing results was ‘‘negative’’, and the results from ELISA should be treated as ‘‘borderline’’ or ‘‘±’’ due to the RASI of 1.5. Interestingly, in the S4 sample, the M-ASI of multiplexing showed pathologic value (2.3) whereas the result from ELISA
Table 1 Comparison of the results (ASI’s and the diagnose-oriented interpretation) in five samples assayed with multiplexing (Multip) and ELISAs in two laboratories, respectively. S1
M R VZV HSV Interpret. ND, not determined.
S2
S3
S4
S5
Multip
ELISA
Multip
ELISA
Multip
ELISA
Multip
ELISA
Multip
ELISA
0.7 0.6 0.9 0.6
0.78 0.7 0.7 0.7
1.2 1.4 1.3 1.2
ND 1.5 1.0 ND ±
0.5 1.1 0.7 0.6
ND 1.2 0.8 0.75
2.3 ND 6.8 0.9 +
1.0 ND 2.1 1.4 +
1.2 1.6 1.9 1.4 +
0.9 2.7 2.6 1.9 +
A. Kułakowska et al. / Methods 56 (2012) 528–531 Table 2 The number of the MRZH-positive cases in the groups investigated. In the MS group, 15/27 (56%) patients showed pathologic ASI’s for at least two antigens. The groups
The ratio of the MRZH positive cases
MS, including: Measles Rubella VZV H. simplex Controls
22/27 (81%) 12/22 (55%) 12/19 (63%) 13/26 (50%) 7/26 (27%) 0/14a
a In one case, a border zone anti-measles specific index of 1.5 was observed; in one case, the indices of antibodies against rubella and VZV were undetectable and in another case, the index of antibodies against rubella was undetectable.
531
range (the same dilution is applied); and (d) it is impossible to repeat analysis of juts one biomarker (or a subset of them) if it has to be re-assayed (due to unsatisfactory quality, for example). In the preparation and validation process of new multiplexing assays, one additional important issue must be considered, namely detection antibodies for all biomarkers of the kit must not cross-react with one another. Summarizing, we think that the multiplexing technology reported in our study could be taken into consideration as a plausible method to detect the polyspecific intrathecal immune reaction in multiple sclerosis. Acknowledgments
incubation time, the same concentration of the detection antibody, all the assay performed on the same day by the same person, etc.) and this is precisely what multiplexing technologies offer: all biomarkers in a given sample are incubated, washed, and analyzed under precisely the same conditions. So far, even distinct differences between the ASI’s still being in ‘‘normal ranges’’ (for example, 0.7 for M and 1.3 for R, which means a two-fold difference) could have been explained either by biological phenomenon (if R-ASI is double as high as M-ASI it might suggest that there is indeed intrathecal immune response against rubella virus, even if the result is still nominally ‘‘normal’’, as otherwise one would have to assume higher diffusion ability of the IgG molecules specific against one antigen than the other) or simply by assay-to-assay differences. It is also perhaps worth mentioning that ‘‘not detected’’ is a plausible and correct result, namely occurring in patients without presence of the specific antibodies against a given antigen in the blood. In such cases, since these non-existing antibodies cannot diffuse into the CSF, it is mathematically impossible to calculate the corresponding index. Therefore, reporting a plausible, numerical value of the ASI provides actually two pieces of information: that a patient’s blood contains specific antibodies, and that they are/are not generated intrathecally. For the antigens considered in this study, it might be of less relevance, as majority of the population have antibodies against these antigens (due to contact with viruses or immunization), however, for example in case of Borrelia burgdorferi, a measurable and plausible index provides very important information, that the specific antibodies are present in the blood and might suggest ongoing systemic infection. On the other hand, multiplexing technology has disadvantages: (a) it requires special equipment, which is still expensive; (b) it is often analytically less sensitive than ELISA; (c) it requires that all analytes in a given sample are within comparable concentration
We are grateful to Drs: H. Möhlig, M. Hartmann, and T. Schumacher (Institut Virion/Serion GmbH, Würzburg, Germany) for very helpful discussions. N.L. was supported by the LLP-ERASMUS Program funds. References [1] H. Reiber, P. Lange, Clin. Chem. 37 (1991) 1153–1160. [2] H. Reiber, S. Ungefehr, C. Jacobi, Mult. Scler. 4 (1998) 111–117. [3] J. Brettschneider, H. Tumani, U. Kiechle, R. Muche, G. Richards, V. Lehmensiek, A.C. Ludolph, M. Otto, PLoS ONE 4 (2009) e7638. [4] P. Lewczuk, J. Wiltfang, Proteomics 8 (2008) 1292–1301. [5] D.A. Vignali, J. Immunol. Methods 243 (2000) 243–255. [6] R.T. Carson, D.A. Vignali, J. Immunol. Methods 227 (1999) 41–52. [7] P. Lewczuk, J. Kornhuber, H. Vanderstichele, E. Vanmechelen, H. Esselmann, M. Bibl, S. Wolf, M. Otto, U. Reulbach, H. Kolsch, F. Jessen, J. Schroder, P. Schonknecht, H. Hampel, O. Peters, E. Weimer, R. Perneczky, H. Jahn, C. Luckhaus, U. Lamla, T. Supprian, J.M. Maler, J. Wiltfang, Neurobiol. Aging 29 (2008) 812–818. [8] M. Zenkel, P. Lewczuk, A. Junemann, F.E. Kruse, G.O. Naumann, U. SchlotzerSchrehardt, Am. J. Pathol. 176 (2010) 2868–2879. [9] O. Hansson, H. Zetterberg, E. Vanmechelen, H. Vanderstichele, U. Andreasson, E. Londos, A. Wallin, L. Minthon, K. Blennow, Neurobiol. Aging (2008). [10] A. Olsson, H. Vanderstichele, N. Andreasen, G. De Meyer, A. Wallin, B. Holmberg, L. Rosengren, E. Vanmechelen, K. Blennow, Clin. Chem. 51 (2005) 336–345. [11] J.F. Kurtzke, Neurology 33 (1983) 1444–1452. [12] W.I. McDonald, A. Compston, G. Edan, D. Goodkin, H.P. Hartung, F.D. Lublin, H.F. McFarland, D.W. Paty, C.H. Polman, S.C. Reingold, M. Sandberg-Wollheim, W. Sibley, A. Thompson, S. van den Noort, B.Y. Weinshenker, J.S. Wolinsky, Ann. Neurol. 50 (2001) 121–127. [13] C.H. Polman, S.C. Reingold, G. Edan, M. Filippi, H.P. Hartung, L. Kappos, F.D. Lublin, L.M. Metz, H.F. McFarland, P.W. O’Connor, M. Sandberg-Wollheim, A.J. Thompson, B.G. Weinshenker, J.S. Wolinsky, Ann. Neurol. 58 (2005) 840–846. [14] M. Andersson, J. Alvarez-Cermeno, G. Bernardi, I. Cogato, P. Fredman, J. Frederiksen, S. Fredrikson, P. Gallo, L.M. Grimaldi, M. Gronning, et al., J. Neurol. Neurosurg. Psychiatry 57 (1994) 897–902. [15] H. Reiber, J.B. Peter, J. Neurol. Sci. 184 (2001) 101–122. [16] H. Reiber, Mult. Scler. 4 (1998) 99–107.
Contents lists available at SciVerse ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
Multiplexing analysis of the polyspecific intrathecal immune response in multiple sclerosis Alina Kułakowska a, Barbara Mroczko b, Maria Mantur c, Natalia Lelental d, Joanna Tarasiuk a, Katarzyna Kapica-Topczewska a, Ute Schulz d, Peter Lange e, Rüdiger Zimmermann d, Johannes Kornhuber d, Piotr Lewczuk d,⇑ a
Department of Neurology, Medical University of Białystok, Poland Department of Biochemical Diagnostics, Medical University of Białystok, Poland Department of Clinical Laboratory Diagnostics, Medical University of Białystok, Poland d Department of Psychiatry and Psychotherapy, Universitätsklinikum Erlangen, Germany e Lab for Clinical Neurochemistry, Department of Neurology, University of Göttingen, Germany b c
a r t i c l e
i n f o
Article history: Available online 16 March 2012 Keywords: Multiple sclerosis Clinical neurochemistry Cerebrospinal fluid MRZH-reaction Multiplexing Neuroimmunology
a b s t r a c t Intrathecal synthesis of the antibodies specific to neurotrofic viruses: measles (M), rubella (R), VaricellaZoster (Z), and/or H. simplex (H), known as ‘‘MRZH-reaction’’ plays important diagnostic role in multiple sclerosis (MS). Whereas the analysis of the oligoclonal IgG bands provides high sensitivity, the MRZHreaction shows high specificity, and hence these methods complement each other. For the first time we applied multiplexing bead-based technology to simultaneously analyze cerebrospinal fluid (CSF) and serum concentrations of antibodies against these viruses, and to calculate the antibody specific indices (ASI’s). The method shows reasonable precision: intra-assay, 2.9–6.7%, and inter-assay, 2.0–3.2%. The results are comparable with these obtained with other methods (ELISAs), including two runs of the certified external quality control schemes. Eighty-one percent of the MS cases (n = 27) and none of the sexand age-matched controls (n = 14), except one subject with ‘‘borderline’’ anti-measles ASI of 1.5, showed intrathecal synthesis of IgG against at least one of the viruses discussed. The ratios of the MRZH-positive cases in the MS group were: 12/22 for M, 12/19 for R, 13/26 for Z, and 7/26 for H. We conclude that the multiplexing technology can be applied as a tool to study the intrathecal immune response in the diagnosis of MS. Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction Twenty years ago, the intrathecal polyspecific immune reaction was described for the first time as a specific diagnostic tool in multiple sclerosis (MS) [1]. The detection of the intrathecal IgG against: measles (M), rubella (R), Varizella-Zoster (Z), and/or H. simplex (H) viruses in MS turned out to be of different quality than the detection of the oligoclonal IgG bands (OCB) on the isoelectrofocusing (IEF): whereas the presence of OCB is very sensitive, and MS subjects without them are extremely rare, it is not specific at all. On the other hand, detection of the intrathecal synthesis of the M, R,
Abbreviations: ASI, antibody specific index; CSF, cerebrospinal fluid; H, Herpes simplex virus; IEF, isoelectric focusing; M, measles virus; MS, multiple sclerosis; OCB, oligoclonal IgG bands; R, rubella virus; Z, Varicella-Zoster virus. ⇑ Corresponding author. Address: Lab for Clinical Neurochemistry and Neurochemical Dementia Diagnostics, Department of Psychiatry and Psychotherapy, Schwabachanlage 6, 91054 Erlangen, Germany. Fax: +49 9131 85 34238. E-mail address: [email protected] (P. Lewczuk). 1046-2023/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymeth.2012.03.002
Z, and/or H viruses, correspondingly frequently referred to as ‘‘MRZ-’’ or ‘‘MRZH-reaction’’, provides specific information already at the moment of the first clinical symptoms [2,3]. The growing number of potentially important biomarkers to be analyzed in a limited volume of the cerebrospinal fluid (CSF) makes techniques of a simultaneous analysis of several parameters in a single small-volume CSF sample a method of choice in the future [4]. Such technologies are generally defined as ‘‘multiplexing’’ independently of the physical–chemical background of the method. For example, the flow cytometric-based Luminex xMAP technology involves coupling of specific capturing ligands to the surface of microsphere (bead) sets uniquely identified with a combination of two fluorescence dyes [5,6]. This allows simultaneous reaction with up to one hundred (theoretical limit) antigens in a single sample. Recently, we [7,8] and others [9,10] applied Luminex-based technology to simultaneously measure several biomarkers of neurodegeneration and neuroinflammation. Following these studies, in this report we applied for the first time a multiplexing technology to simultaneously analyze the intrathecal generation of the
A. Kułakowska et al. / Methods 56 (2012) 528–531
antibodies against M, R, Z, and H viruses in patients with multiple sclerosis. 2. Materials and methods 2.1. The assay and the Luminex platform setting The concentrations of the specific antibodies were measured with the Serion Multianalyt™ bead-based assay (Institut Virion/ Serion GmbH, Würzburg, Germany), which enables simultaneous measurement of the concentrations of IgG specific against measles (M), rubella (R), Varizella-Zoster (VZV), and Herpes simplex (H) viruses in a small volume of a serum or a CSF sample. Serum (diluted 1:100 if not stated otherwise) and CSF (diluted 1:2 if not stated otherwise) samples from a given patient were analyzed (in duplicate) always on the same plate and in the same analytical run. To wash the filter plates, an automated vacuum washer (ELx50, BioTek, Bad Friedrichshall, Germany) was used. Bead populations were sorted, and the signal intensities were measured with the Luminex reader (Austin, USA). Serion Multianalyt™ assays use microspheres (beads) with fluorescence color-coding different than other Luminex-assays. Therefore, the Luminex 100 IS software had to be adjusted to enable raw data acquisition and storage. These data, stored as ‘‘run-files’’, were then analyzed by the Serion Multianalyt™ easyPLEX software. The Luminex 100 IS software was set to analyze at least 100 beads/population with the sample volume set to 60 lL and the sample timeout set to 40 s. 2.2. Linearity: intra- and inter-assay imprecision and external quality control (EQC) To check for the linearity of the measurements, a pooled serum sample was serially diluted with the assay buffer from 1:10 up to 1:1280 with the step of 2 (eight dilution points). To analyze intra-assay imprecision, two human CSF samples (diluted 1:2, and 1:8) and two serum samples (diluted 1:100 and 1:200) were assayed on one plate eight to ten times. The results are presented as the corresponding coefficients of variation (CV’s) of the concentrations of the antigen-specific IgG’s. Inter-assay imprecision was checked by the analysis of the median fluorescence intensity (MFI) of the signal of the positive control sample, included in the assay, measured in six repetitions on six different days. The results are presented as the corresponding CV’s. External (inter-laboratory) quality control was performed by the comparison of the results obtained in our laboratory with the consensus values obtained by n = 16–26 laboratories participating in two runs of the EQC scheme provided by the ISO15189-accredited Society for Promotion of Quality Assurance in Medical Laboratories (Instand, Düsseldorf, Germany). Since our laboratory was the only one participant using the multiplexing platform, our results were compared to, and had to agree with, the consensus values obtained on the ELISA platform. The results of the EQC are presented as the corresponding percentage deviations of our results from these consensus values.
529
2.4. Patients and the routine CSF/serum analyses All the CSF/serum samples were collected at the Department of Neurology, Medical University of Białystok, following the approval of the local ethical committee. All patients gave informed written consent for the study, and all of them underwent the lumbar punctures (LP) for the routine diagnostic procedures. Analyzed were CSF/serum sample pairs from 27 MS (average age, 39.8 ± 13.7 year, 21 females) and 14 sex- and age-matched control patients (average age, 37.6 ± 14.0 year, 11 females). The degree of the neurological impairment in the MS patients was evaluated using the Expanded Disability Status Scale (EDSS, [11]). According to the McDonald criteria [12,13], 21 patients had the relapsing–remitting form of the disease (RR-MS), and six patients had the primary progressive form (PP-MS). None of the MS patients was treated with any diseasemodifying drugs or corticoids at the time of the LP. MRI was performed in all MS subjects. Control group (n = 14) included patients with excluded neuroinflammation, namely: idiopathic headache (n = 6), idiopathic (Bell’s) facial nerve palsy (n = 3), ischialgia due to discopathy (n = 3), and polineuropathy (n = 2). After the LP, the CSF and blood samples were centrifuged (2000g, 20 min), and the supernatants of the CSF and the serum samples were aliquoted, frozen at 80 °C, and shipped to Erlangen on dry ice. IEF was performed with the Hydragel 9 CSF system (Sebia, Fulda, Germany) on agarose gel to fractionate the proteins, followed by immunofixation with peroxidase labelled anti-IgG to detect the OCB. Albumin and IgG concentrations were measured with immune-based nephelometry (Immage, Beckman Coulter) with the CSF and serum samples analyzed always in the same analytical run and correlated to the same calibrator. All MS showed intrathecal synthesis of IgG either on the IEF (the OCB pattern of the type ‘‘2’’ or ‘‘3’’ according to the consensus paper, [14]) or in the quantitative analysis according to Reiber [15]. Control patients showed neither clinical nor neurochemical symptoms of neuroinflammatory disorder; in particular, intrathecal IgG synthesis was not detected in these subjects. 2.5. Calculations of the antibody specific indices (ASI’s) If not stated otherwise, the data are expressed as the ASI’s, calculated according to Reiber [1,16]. Briefly, the concentration of the antibody of a given class (IgG in this study) against a given antigen (M, R, VZV, or H, respectively) in the CSF sample was divided by the concentration of the specific antibody in the corresponding serum sample (after adjusting with the dilution factors). So obtained ‘‘specific IgG quotient’’ was then divided either by the total IgG concentration quotient (QIgG), if no intrathecal synthesis was observed on the corresponding IgG-Reibergram, or by the upper limit of the total IgG concentration quotient (QIgG-Lim), if the intrathecal IgG synthesis was observed on the Reibergram (i.e. when QIgG > QIgG-Lim). ASI’s within the range of 0.6–1.5 were considered ‘‘normal’’, i.e. showing no evidence for the intrathecal IgG synthesis against given virus. The results higher than 1.5 were considered as suggesting the intrathecal specific IgG synthesis against a given antigen, and the results lower than 0.6 were considered implausible.
2.3. Validation of the assay: comparison with the ‘‘classic’’ ELISAs
3. Results
Five CSF/serum sample pairs were assayed with multiplexing in our laboratory, and their fresh aliquots were shipped to the Laboratory for Clinical Neurochemistry, University of Göttingen, where the analyses were re-performed with ELISAs (Dade Behring, Marburg, Germany) according to the routine protocols, and without prior knowledge of the multiplexing results.
3.1. Assay performance: linearity, imprecision, and the EQC results The results of the serum sample dilution experiment are presented in the Fig. 1. The curves show linear performance of the assay at least in the range of the dilution 1:40–1:640. The dilutions 1:10 and 1:20 resulted in the signals above the highest standard
530
A. Kułakowska et al. / Methods 56 (2012) 528–531
was normal (1.0), and in the S5 sample the pathologic H-ASI in the ELISA met the normal value of the multiplexing. In both cases, however, the overall interpretation of the results remained the same independently of the method, due to the agreement of the remaining ASI’s. 3.3. Polyspecific intrathecal immune response in patients; plausibility of the results The results of the MS and control patients are presented in the Table 2. Eighty-one percent of the MS cases and none of the controls (with one ‘‘borderline’’ M-ASI of 1.5 in one subject) showed the intrathecal synthesis of IgG against at least one antigen. In one MS and in one control patient, the ASI were lower than 0.6 (M, and VZV, respectively), otherwise all the results were plausible.
Fig. 1. The dilution curves prepared by a serial dilution of a pooled serum sample. The arbitrary concentrations in the sample diluted 1:10 is defined as ‘‘1000’’. Each point represents the average of double measurements. Dashed vertical line indicates the dilution 1:100, i.e. the usual dilution of a serum sample. MFI, mean fluorescence intensity.
4. Discussion point, whereas the dilution of the serum sample 1:1280 resulted in still measurable signals (i.e. distinguishable from the blank). Correspondingly, the typical dilution, as suggested by the manufacturer and applied in most of the samples of this study (1:100, indicated with a dashed line on the Fig. 1), is in the middle of the linear part of the curve. The CV’s of the intra-assay imprecision were: 4.4–6.7% for measles, 4.0–4.7% for rubella, 2.9–6.4% for VZV, and 4.4–5.9% for H. simplex, with better performance when the CSF sample was diluted 1:2 compared to 1:8, and comparable results when the serum sample was diluted 1:100 or 1:200. The inter-assay CV’s of the MFI of the signals of the control sample assayed on six different plates were: 2.5% for measles, 2.0% for rubella, 2.6% for VZV, and 3.2% for H. simplex. In two runs of the EQC scheme, the deviations of the results obtained in our laboratory from the consensus values were: +8% and 17% for measles, +4% and 17% for rubella, 20% and 16% for VZV, and +4% and 10% for H. simplex. In both runs all the results were found acceptable by the EQC scheme provider, and all the diagnose-oriented interpretations based on the results obtained were correct.
In this report, for the first time we present the application of the multiplexing technology to simultaneously quantify humoral immune reaction against neurotrofic viruses (measles, rubella, Varizella-Zoster, and H. simplex) as a diagnostic tool in multiple sclerosis. This phenomenon, originally described about 20 years ago by Reiber [1] has been since then referred to as ‘‘polyspecific intrathecal immune response’’ or ‘‘MRZ(H) reaction’’. Compared to the original results reported by Reiber [1], we observed almost identical ratios of the MS patients with the intrathecal synthesis of the IgG against: rubella (63% vs. 60%), VZV (50% vs. 55%), and H. simplex (27% vs. 28%), with somehow lower ratio of the MS patients positive for measles (55% vs. 78%). This discrepancy, which resulted in a lower overall ratio of ‘‘positive’’ subjects in our study (81% vs. 90%), might be explained for example by the geographical and generational differences of the populations studied (our MS patients were born about 20 years after these studied by Reiber), resulting in different anti-measles systemic immune responses, for example due to different anti-measles immunization schemes. Simultaneous analysis of many biomarkers with multiplexing technology has several advantages compared to the classic oneparameter methods (like ELISA): (a) it saves precious sample volume, which is particularly important if body fluid difficult to obtain are analyzed (like the CSF); (b) it reduces workload (one run instead of four, for the MRZH-reaction, or even dozens for other applications); (c) it improves the quality control, as it is easier to confirm or to exclude mistakes, like pipetting errors; (d) it improves plausibility control and the diagnosis-oriented interpretation of the results. Particularly the latter point is important for the MRZH reaction: in theory, lack of the intrathecal immunoglobulins synthesis should result, ex definitio, in all ASI’s equal to 1.0, or less conservatively spoken, all the ASI’s in a given CSF/serum sample pair should be more or less the same. This, however, requires the assumption that all the corresponding reactions are performed under precisely the same conditions (the same temperature and
3.2. Comparison of the multiplexing results with the ELISAs The between-methods comparisons of the results are presented in the Table 1. In four out of five sample pairs, not only similar ASI’s were obtained with both methods performed in two different laboratories, but perhaps even more importantly, the diagnose-oriented interpretation (the presence or the absence of the polyspecific intrathecal immune response) was identical: twice ‘‘negative’’ and twice ‘‘positive’’. In one sample (S2), the interpretation of the multiplexing results was ‘‘negative’’, and the results from ELISA should be treated as ‘‘borderline’’ or ‘‘±’’ due to the RASI of 1.5. Interestingly, in the S4 sample, the M-ASI of multiplexing showed pathologic value (2.3) whereas the result from ELISA
Table 1 Comparison of the results (ASI’s and the diagnose-oriented interpretation) in five samples assayed with multiplexing (Multip) and ELISAs in two laboratories, respectively. S1
M R VZV HSV Interpret. ND, not determined.
S2
S3
S4
S5
Multip
ELISA
Multip
ELISA
Multip
ELISA
Multip
ELISA
Multip
ELISA
0.7 0.6 0.9 0.6
0.78 0.7 0.7 0.7
1.2 1.4 1.3 1.2
ND 1.5 1.0 ND ±
0.5 1.1 0.7 0.6
ND 1.2 0.8 0.75
2.3 ND 6.8 0.9 +
1.0 ND 2.1 1.4 +
1.2 1.6 1.9 1.4 +
0.9 2.7 2.6 1.9 +
A. Kułakowska et al. / Methods 56 (2012) 528–531 Table 2 The number of the MRZH-positive cases in the groups investigated. In the MS group, 15/27 (56%) patients showed pathologic ASI’s for at least two antigens. The groups
The ratio of the MRZH positive cases
MS, including: Measles Rubella VZV H. simplex Controls
22/27 (81%) 12/22 (55%) 12/19 (63%) 13/26 (50%) 7/26 (27%) 0/14a
a In one case, a border zone anti-measles specific index of 1.5 was observed; in one case, the indices of antibodies against rubella and VZV were undetectable and in another case, the index of antibodies against rubella was undetectable.
531
range (the same dilution is applied); and (d) it is impossible to repeat analysis of juts one biomarker (or a subset of them) if it has to be re-assayed (due to unsatisfactory quality, for example). In the preparation and validation process of new multiplexing assays, one additional important issue must be considered, namely detection antibodies for all biomarkers of the kit must not cross-react with one another. Summarizing, we think that the multiplexing technology reported in our study could be taken into consideration as a plausible method to detect the polyspecific intrathecal immune reaction in multiple sclerosis. Acknowledgments
incubation time, the same concentration of the detection antibody, all the assay performed on the same day by the same person, etc.) and this is precisely what multiplexing technologies offer: all biomarkers in a given sample are incubated, washed, and analyzed under precisely the same conditions. So far, even distinct differences between the ASI’s still being in ‘‘normal ranges’’ (for example, 0.7 for M and 1.3 for R, which means a two-fold difference) could have been explained either by biological phenomenon (if R-ASI is double as high as M-ASI it might suggest that there is indeed intrathecal immune response against rubella virus, even if the result is still nominally ‘‘normal’’, as otherwise one would have to assume higher diffusion ability of the IgG molecules specific against one antigen than the other) or simply by assay-to-assay differences. It is also perhaps worth mentioning that ‘‘not detected’’ is a plausible and correct result, namely occurring in patients without presence of the specific antibodies against a given antigen in the blood. In such cases, since these non-existing antibodies cannot diffuse into the CSF, it is mathematically impossible to calculate the corresponding index. Therefore, reporting a plausible, numerical value of the ASI provides actually two pieces of information: that a patient’s blood contains specific antibodies, and that they are/are not generated intrathecally. For the antigens considered in this study, it might be of less relevance, as majority of the population have antibodies against these antigens (due to contact with viruses or immunization), however, for example in case of Borrelia burgdorferi, a measurable and plausible index provides very important information, that the specific antibodies are present in the blood and might suggest ongoing systemic infection. On the other hand, multiplexing technology has disadvantages: (a) it requires special equipment, which is still expensive; (b) it is often analytically less sensitive than ELISA; (c) it requires that all analytes in a given sample are within comparable concentration
We are grateful to Drs: H. Möhlig, M. Hartmann, and T. Schumacher (Institut Virion/Serion GmbH, Würzburg, Germany) for very helpful discussions. N.L. was supported by the LLP-ERASMUS Program funds. References [1] H. Reiber, P. Lange, Clin. Chem. 37 (1991) 1153–1160. [2] H. Reiber, S. Ungefehr, C. Jacobi, Mult. Scler. 4 (1998) 111–117. [3] J. Brettschneider, H. Tumani, U. Kiechle, R. Muche, G. Richards, V. Lehmensiek, A.C. Ludolph, M. Otto, PLoS ONE 4 (2009) e7638. [4] P. Lewczuk, J. Wiltfang, Proteomics 8 (2008) 1292–1301. [5] D.A. Vignali, J. Immunol. Methods 243 (2000) 243–255. [6] R.T. Carson, D.A. Vignali, J. Immunol. Methods 227 (1999) 41–52. [7] P. Lewczuk, J. Kornhuber, H. Vanderstichele, E. Vanmechelen, H. Esselmann, M. Bibl, S. Wolf, M. Otto, U. Reulbach, H. Kolsch, F. Jessen, J. Schroder, P. Schonknecht, H. Hampel, O. Peters, E. Weimer, R. Perneczky, H. Jahn, C. Luckhaus, U. Lamla, T. Supprian, J.M. Maler, J. Wiltfang, Neurobiol. Aging 29 (2008) 812–818. [8] M. Zenkel, P. Lewczuk, A. Junemann, F.E. Kruse, G.O. Naumann, U. SchlotzerSchrehardt, Am. J. Pathol. 176 (2010) 2868–2879. [9] O. Hansson, H. Zetterberg, E. Vanmechelen, H. Vanderstichele, U. Andreasson, E. Londos, A. Wallin, L. Minthon, K. Blennow, Neurobiol. Aging (2008). [10] A. Olsson, H. Vanderstichele, N. Andreasen, G. De Meyer, A. Wallin, B. Holmberg, L. Rosengren, E. Vanmechelen, K. Blennow, Clin. Chem. 51 (2005) 336–345. [11] J.F. Kurtzke, Neurology 33 (1983) 1444–1452. [12] W.I. McDonald, A. Compston, G. Edan, D. Goodkin, H.P. Hartung, F.D. Lublin, H.F. McFarland, D.W. Paty, C.H. Polman, S.C. Reingold, M. Sandberg-Wollheim, W. Sibley, A. Thompson, S. van den Noort, B.Y. Weinshenker, J.S. Wolinsky, Ann. Neurol. 50 (2001) 121–127. [13] C.H. Polman, S.C. Reingold, G. Edan, M. Filippi, H.P. Hartung, L. Kappos, F.D. Lublin, L.M. Metz, H.F. McFarland, P.W. O’Connor, M. Sandberg-Wollheim, A.J. Thompson, B.G. Weinshenker, J.S. Wolinsky, Ann. Neurol. 58 (2005) 840–846. [14] M. Andersson, J. Alvarez-Cermeno, G. Bernardi, I. Cogato, P. Fredman, J. Frederiksen, S. Fredrikson, P. Gallo, L.M. Grimaldi, M. Gronning, et al., J. Neurol. Neurosurg. Psychiatry 57 (1994) 897–902. [15] H. Reiber, J.B. Peter, J. Neurol. Sci. 184 (2001) 101–122. [16] H. Reiber, Mult. Scler. 4 (1998) 99–107.