¶Present address: Institute for Viral Preparations, Russian Academy of Medical Sciences, 109 088 Moscow, Russia.
Myelin protein P0-specific IgM producing monoclonal B cell lines were established from polyneuropathy patients with monoclonal gammopathy of undetermined significance (MGUS)
Version of Record online: 5 MAR 2002
Clinical & Experimental Immunology
Volume 127, Issue 2, pages 255–262, February 2002
How to Cite
KVARNSTRÖM, M., SIDOROVA, E., NILSSON, J., EKERFELT, C., VRETHEM, M., SÖDERBERG, O., JOHANSSON, M., ROSÉN, A. and ERNERUDH, J. (2002), Myelin protein P0-specific IgM producing monoclonal B cell lines were established from polyneuropathy patients with monoclonal gammopathy of undetermined significance (MGUS). Clinical & Experimental Immunology, 127: 255–262. doi: 10.1046/j.1365-2249.2002.01739.x
- Issue online: 5 MAR 2002
- Version of Record online: 5 MAR 2002
- monoclonal gammopathy;
- Epstein–Barr virus;
- B cell clones;
- immunomagnetic technique
Monoclonal expansion of B cells and plasma cells, producing antibodies against ‘self’ molecules, can be found not only in different autoimmune diseases, such as peripheral neuropathy (PN), but also in malignancies, such as Waldenström’s macroglobulinaemia and B-type of chronic lymphocytic leukaemia (B-CLL), as well as in precancerous conditions including monoclonal gammopathy of undetermined significance (MGUS). About 50% of patients with PN-MGUS have serum antibodies against peripheral nerve myelin, but the specific role of these antibodies remains uncertain. The aims of the study were to establish, and characterize, myelin-specific B cell clones from peripheral blood of patients with PN-MGUS, by selection of cells bearing specific membrane Ig-receptors for myelin protein P0, using beads coated with P0. P0-coated magnetic beads were used for selection of cells, which subsequently were transformed by Epstein–Barr virus. The specificity of secreted antibodies was tested by ELISA. Two of the clones producing anti-P0 antibodies were selected and expanded. The magnetic selection procedure was repeated and new clones established. The cells were CD5+ positive, although the expression declined in vitro over time. The anti-P0 antibodies were of IgM-λ type. The antibodies belonged to the VH3 gene family with presence of somatic mutations. The IgM reacted with P0 and myelin-associated glycoprotein (MAG), and showed no evidence for polyreactivity, in contrast to other IgM CD5+ clones included in the study as controls. The expanded clones expressed CD80 and HLA-DR, which is compatible with properties of antigen-presenting cells. The immunomagnetic selection technique was successfully used for isolation of antimyelin protein P0-specific clones. The cell lines may provide useful tools in studies of monoclonal gammopathies, leukaemia, and autoimmune diseases, including aspects of antigen-presentation by these cells followed by T cell activation.
Polyneuropathy associated with monoclonal gammopathy of undetermined significance (PN-MGUS) is regarded as an immune-mediated disorder, where an expanded B cell clone secretes monoclonal IgM or IgG/A M-components (for review see [1,2]). About 50% of the patients with M-component of IgM isotype display a binding specificity for peripheral nerve myelin . The target auto-antigen for these antibodies is mainly represented by carbohydrate epitopes shared by glycoproteins, like the myelin protein P0 and myelin-associated glycoprotein (MAG) , and glycolipids [6,7], like the sulphated glucuronic acid epitope present on, for example, sulphated glucuronyl paragloboside (SGPG). A direct pathogenic role of the antibodies against peripheral nerve myelin has been suggested based on findings that demyelination is associated with deposits of monoclonal antibodies and complement . The antibodies are mainly produced by B-1, CD5+ B lymphocytes and they are polyreactive and have low avidity , as a rule. Cross-reactions against bacterial polypeptides from Citrobacter, Proteus and Campylobacter have been demonstrated , suggesting that infections may have triggered the gammopathy.
The specific role of antibodies against peripheral nerve myelin remains uncertain, however, based on the following observations: (i) There is no clear correlation between the occurrence of anti-MAG antibodies and the type or severity of disease . (ii) Anti-myelin antibodies may occur in healthy blood donors . (iii) There is no distinct correlation between the decrease of antiperipheral nerve myelin antibody level and clinical effect upon immunosuppressive treatment . (iv) Many patients with PN-MGUS without antimyelin antibodies may still respond to immunosuppressive treatment . Thus other mechanisms, besides the IgM monoclonal antibodies, may be involved in the pathogenic process. In particular, several reports point at a participation of T cells [15–18], although the putative mechanisms so far are unclear.
The establishment of B cell lines and clones would provide a useful tool to study further the role and biological functions of autoimmune myelin-specific B cells and would also facilitate studies on B–T-cell interactions in the pathogenesis of PN-MGUS. Epstein–Barr virus (EBV) transformation of B cells, as a method [19,20], has been used to obtain autoantibody-producing B cell lines in a number of autoimmune diseases, such as systemic lupus erythematosus , myasthenia gravis , multiple sclerosis , and autoimmune thyroiditis [24,25]. In MGUS, where a clone of B cells already exists in vivo, the experience of establishing B cell clones is limited. Andersson et al. used EBV transformation to reveal and establish anti-idiotypic B cells. However, no attempts have been made to use EBV transformation to study myelin-specific B cells in patients with PN-MGUS. Instead, human × human hybridomas producing anti-MAG antibodies were derived from fusion of MGUS patient’s blood cells with the UC lymphoblastoid cells . No genetic abnormalities related to the MGUS condition were revealed. This was, however, studied only at the chromosomal level. In a parallel system of clonal B cell expansion, in which autoantibodies may occur, we have immortalized the malignant clones of several B-type chronic lymphocytic leukaemia patients and compared these to their normal counterparts .
In the transformation process of autoantibody-producing cells, it would be of advantage to preselect B cells with the desired specificity. Biotinylated autoantigens and subsequent fluorescence activated cell sorting showed a substantial enrichment of antigen-specific cells . Immunomagnetic technique has recently been successfully used by our group for the same purpose . The aim of the present study was to establish a feasible technique to establish B cell lines from patients with MGUS, utilizing immunomagnetic enrichment of myelin-specific B cells followed by EBV-transformation. Phenotypic and genomic characterization is shown for B cell lines from a patient with PN-MGUS.
MATERIALS AND METHODS
P0-specific B cells were isolated from peripheral blood from two PN-MGUS patients, a 71-year-old woman (TJ) and a 61-year-old man (RG) with PN-MGUS. Patient no. 1 (TJ) had chronic progressive sensory-motor polyneuropathy. The M-component was 5 g/l and of IgM-λ type. Her serum antibodies reacted with crude myelin (medium level), P0-protein (medium level) and MAG (medium level) as measured by ELISA [13,31,32], as well as with the LK-1 glycolipid (high titre) . Patient no. 2 (RG) had chronic progressive sensory-motor polyneuropathy. The M-component was 7 g/l and of IgM-κ type. The serum antibodies displayed a similar broad reactivity to crude myelin (high), P0 (high), MAG (high) and LK-1 (medium high titre). Based on the findings of a broad reactivity to glycoproteins and glycolipids, it is clear that the serum antibodies, from both patients, reacted against carbohydrate epitopes shared by P0, MAG and LK-1. P0 is the most abundant protein in peripheral nerves, and was further used in the experiments.
The patients had given consent to the blood donation, and the local ethics committee at University Hospital of Linköping had granted permission.
For phenotypic FACS-characterization of the B cell clones, the following MoAbs were used: anti-CD3, -CD5, -CD6, -CD14, -CD19, -CD20, -CD22, -CD23, -CD25, -CD56, -CD69, -CD80, -CD95, -HLA-DR. Isotype-matched mouse Ig was used for each MoAb. The MoAb and the isotype controls were purchased from Becton-Dickinson (BD), Dakopatts (Dako), or Coulter.
Lipopolysaccharide (LPS) of E. coli (Sigma Chemical Co., St. Louis, MO, USA); phorbol 12-myristate-14-acetate (PMA) (Sigma); pneumococcal polysaccharide (S3) was a kind gift from Dr Philip Baker (USA); polyvinylpyrrolidon (PVP) with a MW of 350 kDa (Serva, Germany); single-stranded DNA (ssDNA) from sperm whale (Sigma); phosphatidylcholine (PtC) (Sigma); staphylococcal protein A (SpA) (Sigma); human IgG (Kabi, Stockholm, Sweden); and myelin protein P0 were used as the antigens. Preparation of P0 was performed in the following way. First, peripheral nerve myelin was isolated from bovine lumbosacral plexus and prepared according to Kadlubowski et al.. Then, P0 protein was isolated as follows. The myelin was homogenized in 0·03 M HCl containing 1 mM 2-mercaptoethanol (2-ME) and then stirred for 16 h at 4°C. The mixture was centrifuged for 45 min at 96 000 g. The pellet was washed once with 0·03 M HCl containing 1 mM 2-ME and centrifuged at 96 000 g for 30 min. The remaining pellet was homogenized in an ice-cold 2 : 1 mixture of chloroform and methanol. pH was adjusted to 7·0 and the solution was centrifuged for 30 min at 100 000 g. The chloroform : methanol phase separation step was repeated once. The pellet was then dried under nitrogen and dissolved at 100°C in 0·01 M Tris-HCl (pH 7·5) containing 10% (w/v) SDS, 5 mM EDTA, 1 mM dithiothreitol (DTT) and 0·02% (w/v) NaN3. The insoluble material was removed by centrifugation and the supernatant was applied on a HiLoad 16/60 Superdex 75 column (Pharmacia) equilibrated with 0·01 M Tris-HCl (pH 7·5) containing 0·5% (w/v) SDS, 5 mM EDTA, 1 mM DTT and 0·02% (w/v) NaN3. The column was eluted with the same buffer and 2·5 ml fractions were collected. The fractions were analysed with SDS-PAGE. The fractions containing P0 (28 kDa) were pooled and applied to the column again. The fractions containing P0, identified by SDS-PAGE, were pooled and dialysed in 0·02% NaN3 at room temperature for 3 days, and in water at 4°C for another 2 days and thereafter lyophilized. SDS-PAGE was performed in order to control the purity of the P0 preparation. No band of 22 kDa corresponding to PMP22 was found in the preparation.
Conjugation of P0 glycoprotein with FITC
One mg P0 glycoprotein was dissolved in 0·5 ml carbonate-bicarbonate buffer, pH 9·5, and mixed with 40 μg of FITC isomer I (Sigma) in 4 μl DMSO (Sigma) and incubated in the dark on a rocker platform for 1 h. FITC-conjugated P0 was isolated by gel filtration using a PD-10 column (Pharmacia, Uppsala, Sweden). The conjugated P0 was eluted in Tris-buffer, pH 8·2.
The lymphoblastoid cell line (LCL) M2D6, producing monoclonal human IgM, was used as a control cell line. It was obtained from human tonsil B cells using magnetic beads coated by synthetic viral peptide MN-24 (residues 302–322 of V3 loop of gp120 of HIV-1) . Cells were cultivated in Iscove-OptiMem medium (1 : 1) with 10% FCS and antibiotics (Gibco BRL, Life Technology, UK).
Immunomagnetic selection of P0-specific B cells and expansion of clones
Magnetic beads (Dynabeads; Dynal A/S, Oslo, Norway) were conjugated with P0 according to the manufacturer’s instruction using borate buffer and 48 h incubation. Conjugated beads were stored in PBS/BSA, pH 7·4, at +4°C.
The peripheral blood mononuclear cells (PBMC) were isolated from whole blood using Lymphoprep (Nycomed Pharma A/S, Oslo, Norway). The cells were resuspended in Iscove’s modification of Dulbecco’s medium (IMDM) with additives (5% FCS, glutamine, penicillin, streptomycin, and essential amino acids) and depleted for macrophages by plastic adherence. The beads were washed in IMDM medium with additives immediately before use. 38 × 106 PBMC were rotated with 2·9 × 107 beads for 2 h at + 4°C and the rosetted cells were separated with magnetic device and washed twice in IMDM-medium. Remaining 0·8 × 106 PBMC were infected with 1 ml Epstein–Barr virus, strain B95-8 according to previously described method  for 2 h at 37°C. Cells were cultured in Iscove-OptiMem (1 : 1) medium with additives (10% FCS, PEST and glutamine) in round-bottomed plates.
ELISA for determination of Ab in the cell supernatants
Human IgM was analyzed in ELISA by coating Costar Maxisorb flat-bottomed 96-well plates with mouse MoAb to μ-chains of human IgM. 7·5–1000 ng/ml of purified polyclonal human serum-derived IgM (Dako) was used for the standard-curve. Each sample was determined in triplicates. The indicator Ab for human IgM ELISA was either HRP-anti-IgM MoAb (a kind gift from Dr T. Borisova, Moscow) or ALP-anti-μ MoAb (Dako). The substrates were ABTS (Boehring Mannheim GmbH, Germany) or p-nitrophenyl phosphate (Sigma), for the HRP or ALP conjugates, respectively.
Anti-myelin antibodies were detected as previously described . Serum from the TJ and RG patients were used at 1 : 100 dilution as positive Ab controls and supernatants from M2D6 cells (1 : 2 dilution) were used as controls.
ELISA for the polyvalent T-independent antigens LPS, PVP, S3, PtC, SpA, was done with MaxiSorb plates as described above, coated with 100 μl of the antigen (5 μg/ml). The specific Ab in the culture supernatant was always compared to the negative controls in relation to its IgM concentration. Supernatants were diluted 1/2, 1/4 and 1/8 in phosphate buffer saline (PBS) with 0·1% Tween 20 and 0·5% bovine serum albumin (BSA).
Anti-MAG ELISA (Bühlmann, Allschwil, Switzerland) was done on culture supernatants and purified IgM from TJ clones, according to instructions from the manufacturer. IgM was purified by high trap IgM purification column (Pharmacia, Uppsala, Sweden).
FACS phenotyping of B cell clones derived from TJ patient
The phenotypes of TJ-derived B cell clones were determined by flow cytometry. Briefly, cells (1 × 106) were washed once with PBS-2% FCS and incubated with fluorochrome-labelled antibodies for 30 min at 4°C. After the incubation, cells were washed three times with the same buffer and the fluorescence intensity and distribution of 10 000 live cells was performed using a FACSCalibur flow cytometer (Becton Dickinson).
Immunofluorescence of surface immunoglobulins
To determine the surface IgM isotype and kappa/lambda light chains, direct FITC conjugates for IgM-kappa, -lambda, IgD, IgG were used in immunofluorescence. The cells were observed under a Nikon fluorescence-microscope.
Detection of antibacterial and antinuclear antibodies
Routine immunofluorescence methods were used for screening of antibacterial and antinuclear Ab at the Clinical Microbiology laboratory and Clinical Immunology laboratory of Linköping University Hospital, respectively. A panel of 15 of the most frequently isolated bacterial strains was used for the antibacterial screening. The antinuclear antibody testing included immunofluorescence screening on Hep2 cells of the TJ cell supernatant derived Ab, and ELISA screening was performed for SS-A, SS-B, Scl-70, Sm/RNP, and Jo-1.
Determination of VH immunoglobulin gene family and somatic mutation
The rearrangement of VH gene coding anti-P0 antibody was determined in two TJ clones taken from different occasions by PCR-amplification of cDNA using primers specific for the different VH families’. The PCR-product was purified and sequenced in an ALF DNA-sequence instrument (Pharmacia), using the same primers.
Two P0-specific clones, TJ1 and TJ2, were established from the PN-MGUS patients and were maintained in long-term cultures for this study: The TJ1 B cell line secreted IgM-λ antibodies for approximately 6 months, and the TJ2 secreted IgM-λ in a stable fashion until now (two years). The peripheral blood mononuclear cells were isolated by P0-coated magnetic beads, followed by EBV-transformation in order to obtain stable long-term cultures of the antigen-specific B-lymphocytes. The cells were seeded in 96-well microtitre plates and after two weeks of cultivation practically all cultures were positive for antimyelin protein activity as measured by ELISA. After three weeks cells were collected and the selection with P0-coated magnetic beads was repeated. Positive clones were chosen and expanded in flasks. One clone from each long-term experiment was selected for detailed studies and named TJ1 and TJ2, respectively. B cell clones from the RG patient secreted anti-P0 Ab for two months after viral transformation, but failed to proliferate under the culture conditions used in this study.
The M-component observed by electrophoresis in the TJ patient plasma was of IgM-λ type. Investigation of the B cell clones TJ1 and TJ2 showed that these cells exclusively expressed IgM-λ, as revealed by membrane immunofluorescence technique. Figure 1 shows the staining of Ig-light chains where all cells were lambda positive. None of the cells stained for kappa-light chains, and the isotypic control IgG was negative.
In order to determine the frequency of specific anti-P0 B-lymphocytes present in the circulation of the TJ patient, we also stained freshly isolated peripheral blood lymphocytes after blood sampling by two-colour flow cytometry: the pan-B cell marker CD19-PE was used in combination with the specifically labelled antigen FITC-P0. The gating was first performed for small lymphocytes (Fig. 2, upper left FACS-diagram). The lymphocytes were then gated for the pan-B cell marker CD19 (FL2-H). As seen in Fig. 2, lower left panel, B cells expressing the specific anti-P0 B cell surface receptors were clearly observed by this two-colour method. Upper right diagram also shows the frequency of P0-stained cells. The lower right diagram shows the isotype control. The frequency of the autoimmune B cell clone present in circulation was estimated to 0·04% of the total mononuclear cells. The frequency of anti-P0 B cells among CD19+ B cells was 0·36%.
FACS-based phenotypings of TJ1 and TJ2 clones were performed. The results presented in Table 1, taken together with the fact that both Abs were of IgM-λ type with P0-binding specificity, strongly indicate that the two clones are identical. The clones expressed the B cell markers CD19, CD20, CD22, but not T cell marker, CD3, nor the NK cell marker CD56, nor the monocyte marker CD14. Interestingly, CD80 and HLA-DR, which are prerequisite receptors for antigen-presentation, were expressed (Fig. 3). CD5, CD23 and CD25 were expressed on a fraction of the B cell culture (i.e. bi-phasic expression), which could indicate a tendency to divergence with regard to these markers. The cultured B cell clones, TJ1 and TJ2 cells were partly CD5 positive at the initial stage of cultivation, but gradually the CD5-expression declined. Thus after 4 months of cultivation, the number of CD5+ cells was reduced from 35% (data not shown) to 3% (Table 1). Similar loss of CD5 expression was observed during cultivation of monoclonal human M2D6 and E2C5 lymphoblastoid cell lines (E.S., J.N. and A.R., unpublished observations).
|Surface markers||TJ1 clone % positive cells||TJ2 clone % positive cells|
The IgM Ab released from the TJ1 and TJ2 B cells were investigated with regard to their possible cross-reactivity, since it is a well-known observation that IgM derived from B-1, CD5+ cells produce Ig of low avidity and multireactive binding properties. The cell supernatants were found to be negative for reactivity against a panel of the 15 most frequently isolated clinical strains of bacterial antigens (data not shown), as well as a panel of nuclear auto-antigens SS-A, SS-B, Scl-70, Sm/RNP and Jo-1 (data not shown). In addition, ELISA was used for testing of reactivity against another set of multivalent antigens (Table 2). Our previous observation of the CD5+ M2D6 cell line, that produced IgM against HIV-associated V3-loop peptide, revealed a multireactive binding pattern , also obvious in this study (Table 2). The IgM containing supernatant from M2D6 thus served as a suitable control in the present study. The M2D6 antibodies reacted with most antigens used, while the supernatants from TJ1 and TJ2 reacted with myelin antigens exclusively (Table 2). The B cell superantigen SpA, which is known to bind human IgM of certain VH-families, reacted with both M2D6 and TJ supernatants. SpA is a B cell superantigen binding mainly Ig coded by VH3 gene family . In addition to myelin reactivity, the TJ supernatant and purified supernatant IgM were tested for anti-MAG by ELISA, and were found to be strongly positive. The OD values were 1·6–1·7 for supernatants and purified IgM, the highest point of the standard curve was 1·5, a medium control was 0·05, and control IgM was 0·1.
|ELISA value (OD 405 nm)|
|TJ2 IgM||M2D6 (control) IgM|
|Antigens||Exp #1||Exp #2||Exp #1||Exp #2|
The sequence determination of the IgM VH gene of the TJ2 clone showed that it belonged to the VH3-15 gene family (Fig. 4). 8·1% mutations were found, five silent and 19 replacements; 11 in CDR and 13 in FWR; 11 in hot-spots were present. The TJ1 clone showed identical sequence in the VH region including identical mutations.
Autoantibodies are associated with the development of pathological processes in different autoimmune diseases including neurological diseases. About 50% of patients with PN-MGUS have serum antibodies against peripheral nerve myelin, but the specific role of these antibodies remains uncertain. The expanded B cell clone in MGUS does not progress to plasma or malignant (myeloma) cells, as a rule, although exceptions have been reported. The TJ B cell clones were phenotypically similar to CD5+ B-CLL cells . Evidently, the establishment of cell lines secreting anti-P0 autoantibodies may be helpful to study the biological role of these cells and autoantibodies produced by them. The availability of monoclonal antibodies would also permit to study idiotypic regulation/dysregulation in PN-MGUS.
The results presented in this study confirm and expand our previous data [30,34,39], and show that positive selection of B cells bearing antigen-specific receptors with subsequent transformation of selected cells by Epstein–Barr virus should be a most suitable method to establish human B cell lines and to obtain human monoclonal antibodies. An alternative method would be to perform a fusion of B cells with a stable fusion partner . The foremost advantage of the EBV method versus human hybridoma would be a better preservation of the original B cell phenotype.
The TJ clones presented in this study were found to be of monoclonal IgM-λ type, with binding restricted to myelin proteins P0 and MAG, in contrast to M2D6 derived from CD5+ monoclonal cell line known to exhibit a broad reactivity pattern against T independent antigens. The reactivity patterns of the supernatants were identical with the serum M-component. Thus, the two TJ cultures and antibodies were of clonal origin, presumably directed against carbohydrate epitopes shared by P0 and MAG.
The TJ1 and TJ2 clones obtained in this study were phenotypically very similar, with subtle differences only, in the strength of the CD-marker expression (Table 1). This may indicate that these two clones originated from the same clone but at different stages of differentiation. The identical VH3-15 sequence indeed shows that the clones had the same B cell origin. It should be noted that TJ2 cell culture showed a heterogeneous expression of CD5, CD23 and CD25 markers. Some cells were positive while others were negative. The reason may be technical and reflect the sensitivity of detection of the method, but may also indicate a true heterogeneity. During cultivation, TJ1 and TJ2 cells declined in CD5-expression. The same loss was observed previously in our laboratory after prolonged cultivation of M2D6 and E2C5 human B cell clones (not shown). Although it could be an effect of the EBV transformation, the decline in CD5 expression indeed suggests that CD5 is rather an activation marker than a lineage marker. These in vitro data confirm results obtained in our previous study , in which we investigated CD5 expression on B cells of patients with PN-MGUS and concluded that CD5 expressed on B cells may be an activation marker, findings in line with Zupo et al.. Attempts to activate TJ2 cells and induce re-expression of CD5 cells with different agents including antigen-P0 in soluble and in insoluble forms, previously used for the selection of the clone, did not succeed. Even such a highly efficient cell activator as PMA did not induce the CD5 expression. Interestingly, the B cell clones strongly expressed HLA-DR and CD80, compatible with antigen-presenting properties. Although we cannot exclude that part of this expression could be due to the EBV-transformation, we hypothesize that the B cell clones in vivo might display antigen-presenting capacity, which would be a way to activate auto-reactive T-cells. T-cells would then, in turn, among other things, influence the development of the B cell clone. In fact, the findings in the present study of a mutated IgM VH3-15 gene sequence, together with a lack of poly reactivity, indicate that the B cell clone was under the influence of T-helper cells, which is also in line with previous findings [15–18]. Previous studies on antimyelin-associated glycoprotein (MAG) and antinerve glycolipid V-gene sequences revealed a diversity of the immune repertoire and indicated the lack of a public idiotope in contrast to autoantibodies such as rheumatoid factors and cold agglutinins , which is in line with our results presented in this study.
Our present study shows that it is possible to establish stable B cell clones producing autoantibodies to myelin proteins, which now may allow for studies on B cell interaction with T cells, antigen-presentation and regulatory mechanisms in PN-MGUS.
We thank Prof. Urban Forsum at the Laboratory of Clinical Microbiology at Linköping University Hospital for the help with screening Ab for antibacterial activity. We also thank Karin Backteman and Jeanette Svartz at the Laboratory of Clinical Immunology, for excellent help with the phenotyping of clones.
This work was supported by grants from the ‘Network for Inflammation Research’/Swedish Foundation for Strategic Research, the County Council of Östergötland, the Society for Neurologically Disabled (NHR), and the Swedish Cancer Society (2357-B00–15XCC) and (3171-B00–12XAQC). GSD.
- 1Polyneuropathy associated with monoclonal gammopathy. A clinical, neurophysiological and immunological study. Linköping University Medical Dissertation, no. 409, 1994.
- 2Clinical features, evaluation, and treatment of patients with polyneuropathy associated with monoclonal gammopathy of undetermined significance (MGUS). J Clin Apheresis 1999; 14:149–53.DOI: 10.1002/(sici)1098-1101(1999)14:3<149::aid-jca8>3.0.co;2-9
- 19Polyclonal Ig production after Epstein–Barr virus infection of human lymphocytes in vitro. Nature 1977; 26:752–4., , , ,
- 23Human B cell lines secreting IgM antibody specific for myelin basic protein. J Neuroimmunol 1989; 1–2:9–16., , , ,
- 24A study of human–human hybridomas from patients with autoimmune thyroid disease. J Clin Immunol 1987; 1:71–7.,
- 34Human monoclonal antibodies to viral peptides. Human Antibodies 1997; 8:67–9., , ,
- 36Human monoclonal antibodies to MN-24 peptide of gp120 HIV-1. Human Antibodies 1999; 9:107–10.