Journal of Autoimmunity (1992) $771-785
Characterization of Human-Human Hybridoma Monoclonal Anti-Ro( SS-A) Autoantibodies Derived from Normal Tonsil Lymphoid Cells
Helene Massicotte, John B. Harley* and David A. Bell University University
of Western Ontario, Rheumatic Hospital,
Program, University
London,
Oklahoma
Ontario,
Medical
Disease Unit, Department
Canada and *The Arthritis
Research
Foundation,
Department
of Medicine,
and Immunology of Medicine,
of Oklahoma Health Sciences Center and Veterans Administration Medical Center, Oklahoma City, OK, USA (Received 12 February 1992 and accepted 3 August 1992)
Human-human hybridomas obtained from the separate fusion of tonsillar lymphoid cells from three different normal individuals to the lymphoblastoid cell line GM 4672 were screened by ELISA for the presence of autoantibody to Ro(SS-A). Those anti-Ro(SS-A) reactive hybridomas were then cloned by limiting dilution. Nineteen monoclonal IgM antiRo(SS-A) antibodies were obtained, which showed specificity to Ro(SS-A) by ELISA and Western blotting (60 Mla). Some of these monoclonal antiRo(SS-A) antibodies showed reactivity to DNA (2/19), cardiolipin (9/19), Sm/RNP (15/19) by ELISA, and to IgG (12/19) and La(SS-B) (19/19) by ELISA and Western blotting. None showed reactivity to the unrelated proteins caseine and BSA, nor to RNA. Inhibition studies revealed that the binding to Ro( SS-A) of both IgM hybridoma monoclonal and SLE serum polyclonal IgM anti-Ro(SS-A) antibodies was inhibited with Ro(SS-A), La( SS-B) and Sm/RNP but not with IgG, DNA, RNA and BSA. These data indicate that (1) normal humans have the genetic potential to express antibodies to Ro(SS-A) and (2) the normally derived monoclonal and SLE serum IgM anti-Ro(SS-A) antibodies share similar antigen binding properties and therefore may possibly originate from a common pool of precursor B cells.
Introduction
Systemic lupus erythematosus (SLE) is characterized diverse autoantibodies. Among these, autoantibodies
by the presence of many to nucleic acid antigens
Correspondence to: Dr David A. Bell, University Hospital, Rheumatic Disease Unit, Box 5339, London, Ontario, Canada, N6A 5A5. 771 0896-841 l/92/060771
+ 15 $08.00/O
0 1992 Academic Press Limited
772 Helene Massicotte et al. (DNA/histone complex) have been well studied because of their observed relationship to and association with, tissue injury and disease activity. More recently, increasing interest has focused on a variety of other autoantibodies directed against RNA-protein complexes e.g. Ro(SS-A), La(SS-B), Sm, UlRNP, whose presence identifies certain subsets of disease and/or related connective tissue syndromes such as SLE [l] and Sjogren’s syndrome [2] or occurs in mothers of infants with neonatal SLE or congenital heart block [3]. The investigation of these latter autoantibodies has accelerated recently with the development of greater insights into the molecular nature of the target autoantigens of these RNA-protein complexes. Briefly, it has been demonstrated that Ro(SS-A), a constituent of many normal tissue cells [P61, is present on a particle comprised of RNA of 83-112 bases (hyl, hy2, hy4 and hy5 in man) [7,8] and one or more associated polypeptides (60,54 and/or 52 kDa, depending on the cell type) [9, lo] which bear the antigenic epitopes [ 1 l-l 31. La( SS-B), another eukaryotic small RNP particle, contains RNA of about 80-l 30 nucleotides including precursors of RNA polymerase III transcripts [ 14,151, Ul RNA [ 161, certain small viral RNAs [ 171 and a highly conserved 43-50 kDa phosphoprotein [18, 191. Ro(SS-A) may exist alone or complexed to La(SS-B) depending on the YRNA present [2,11-12,20-241. The ability spontaneously to generate human hybridoma monoclonal autoantibodies to other nuclear antigens by the fusion of tonsillar lymphocytes of a normal individual to the human lymphoblastoid cell line GM 4672 [25,26], and the observation of low levels of anti-Ro(SS-A) antibodies in normal donor plasma [27], raised the possibility that natural antibodies to Ro(SS-A) could be produced in vitro. The present paper shows that autoantibodies to Ro(SS-A) can be successfully obtained by this strategy of somatic cell hybridization with normal human lymphocytes and explores the reactivity and specificity of these anti-Ro(SS-A) antibodies with other RNA-protein complexes, DNA, IgG and cardiolipin.
Materials and methods 1. Production of human :human hybridomas The production and propagation of humanhuman hybridoma autoantibodies using normal whole tonsillar lymphoid cells fused to the human lymphoblastoid cell line GM 4672 was performed as previously described [26]. The tonsils used were from normal children who did not demonstrate anti-Ro(SS-A) antibodies in their serum and had no family history of connective tissue disorder.
2. ELISA
autoantibody
screening
Both IgM and IgG autoantibody production were screened by ELISA in undiluted supernatants of all hybridomas and their clones, using alkaline phosphatase conjugates, F(ab’), goat anti-human IgM (u chain specific) and F(ab’), goat anti-human IgG (y chain specific) respectively (Jackson Immunoresearch Laboratory, Mississauga, Ont, Canada). The ELISA used for the detection of antibodies to ssDNA, DNA, RNA and cardiolipin were described previously [25,26]. Standard sera of KAPPS were used as controls for the cardiolipin assay.
Normal human monoclonal anti-Ro(SS-A)
antibody
773
The affinity-purified calf thymus or spleen Ro(SS-A) [2, 13,27,28], the afBnitypurified bovine thymus or spleen La(SS-B) and Sm/RNP (Immunovision, Springdale, AR) were coated overnight at 4°C in carbonate coating buffer, pH 9.6 at 3 pg/ml. The supernatants as well as the conjugate were both subsequently incubated on the plates overnight at 4°C. The washing buffer contained O.lq/, bovine serum albumin (BSA) (Boehringer Mannheim, Gmbh, Germany) in 0.1 M Tris-HCl, 0.059b Tween 20, pH 7.4 (Fisher Scientific Co, Fair Lawn, NJ, USA). The binding of hybridoma IgM antibodies to human IgG (rheumatoid factor ELISA assay) used whole human IgG (Cappel, Organon Teknika Inc, Scarborough, Ont., Canada), Fc or Fab’ fragments of IgG (Jackson) coated at 10 pg/ml in carbonate coating buffer. A diluted (in 1 o/0bovine serum albumin (BSA)-saline) rheumatoid factor-positive serum control or undiluted supernatants were added to antigen coated plates, previously blocked with 3% BSA-saline, and incubated overnight at 4°C. F(ab’), goat anti-IgM alkaline phosphatase conjugate (Jackson) was used at a dilution of 1:2,500 in 30/6BSA-0.05% Tween-salineandleft for 3 hat roomtemperature (RT). All anti-Ro(SS-A) antibody cloned supernatants were also tested for reactivity to the unrelated antigens BSA and caseine (BDH Chemicals Ltd, Poole, UK), coated at 40 ug/ml in carbonate coating buffer or to wells without antigen. 3. Sera SLE sera (5) with IgG anti-Ro(SS-A) antibody reactivity were studied in order to compare their IgM and IgG ELISA binding to Ro( SS-A), La( SS-B) and Sm/RNP, and to perform inhibition assays. The results of two representative sera, BRA and ASH, are shown in Figure 3. ASH serum by ELISA had IgM and IgG antibodies to Ro(SS-A), La(SS-B) and Sm/RNP; BRA serum had IgM antibodies to all three antigens and IgG antibodies to Ro(SS-A) and La(SS-B) only. 4. Inhibition of ELISA
Ro (SS-A)
binding
Inhibition assays were performed to confirm the Ro(SS-A) specificity of the cloned antibodies. Anti-Ro(SS-A) antibodies from sera or from cloned supernatants were premixed in a test tube with either phosphate buffered saline (PBS; 100% binding value), Ro(SS-A), La(SS-B), Sm/RNP, RNA, BSA, ssDNA, DNA, or whole IgG at 25, 5 and 1 gg for IgM antibody inhibition or 15,5, 1 and 0.1 gg for IgG antibody inhibition. The mixture was incubated overnight at 4°C then plated on Ro(SS-A)coated plates previously blocked with hybridoma growth medium (HGM) for 2 h at RT. The mixture was left on the plates for 3 h at 4°C then washed off and the conjugate added as previously described. The percent binding was calculated as follows: [(OD405llln+ inhibitor)/(OD),,nm + PBS)] x 100. 5. Heavy and light chain class expression of hybridoma monoclonal antibodies To determine and quantify the heavy chain (l.tor y chain) expressed in the hybridoma supernatants, either F(ab’), goat anti-human IgM or IgG (heavy chain-specific, Jackson) were coated in 10 pg/ml carbonate coating buffer. Serially diluted
774
Helene Massicotte et al.
supernatants as well as IgM and IgG standards (Cappel, range 0.01 to 2 p,g/ml) were incubated on the coated plates overnight at 4°C. The conjugates used were, respectively, a goat anti-human IgM (Zymed Laboratories, Mississauga, Ont., Canada) or goat anti-human IgG (Kirkergaard Laboratories, Mandel Scientific Company Ltd, Guelph, Ont., Canada) alkaline phosphatase diluted 1:2,000 in 0.1 y0 BSA, 0.1 M Tris-HCl pH 7.4 and incubated for 3 h at RT. The curves established with the standards were used to determine the concentrations of IgM or IgG in the hybridoma supernatants. The light chains expressed by and associated with the IgM antibodies produced in the hybridoma supernatants were determined in a similar fashion. F(ab’), goat anti-human IgM (Jackson) was coated as above. A goat anti-human lambda (h) or kappa (K) chain-specific alkaline phosphatase conjugate (Sigma) diluted at 1:6,000 in 1 Y0 BSA, 2% bovine gamma globulin (BGG), 0.1 M Tris-HCl pH 7.4 (BGG, Cohn fraction II, III, Sigma) was used to develop this assay. The GM 4672 supernatant (IgG K) was used as a negative control and two Waldenstrom sera (IgM h or IgM K) diluted 1: 1,000 were employed as positive controls for h and K light chain assays, respectively. 6. SDS-PAGE Ro(SS-A), La(SS-B), Sm/RNP, IgG (whole, Fc and Fab’) and normal human lymphocyte lysates were separated by SDS-PAGE (sodium dodecyl sulfatepolyacrylamide gel electrophoresis). These antigens were reduced by boiling for 2 min at 96°C in 5% p-2 mercaptoethanol (Sigma), and electrophoresed according to Laemmli [29] with 15% and 5% polyacrylamide separating and stacking gels, respectively, both with 0.1% SDS. The SDS-PAGE was performed in a MiniProtean II apparatus (BioRad Chemical Division, Mississauga, Ont., Canada) according to the manufacturer. Prestained molecular weight markers (Gibco/BRL, Burlington, Ont., Canada) were run on the gel along with the antigens. 7. Immunoblots
The proteins separated by SDS-PAGE were then transferred electrophoretically from the gel to nitrocellulose (BioRad) [30] in a Mini-Trans Blot cell apparatus (BioRad) according to the manufacturer, at 100 V for 1 h with stirring and cooling. The nitrocellulose cut into strips was blocked overnight at 4°C with 5% powdered skim milk in Tris-buffered saline (TBS). Next, the strips were washed with TBS and incubated with undiluted cloned supernatants or control positive SLE or normal sera for 3 h at RT. After extensive washing, the F(ab’), goat anti-human IgM alkaline phosphatase conjugate (Jackson) at 1:2,500 in 1 96 BSA-TBS was added for a further incubation of 2 h at RT. The strips were again washed and then developed with NBT/BCIP substrate (Gibco/BRL) and the reaction stopped by water washes, and read. 8. Immunoprecipitation The immunoprecipitation, electrophoresis and silver staining of the RNAs immunoprecipitated by 10 of the monoclonal anti-Ro(SS-A) antibodies was performed according to Forman [31] with the following modification. Protein A beads were
Normal human monoclonal anti-Ro(SS-A)
antibody
775
precoated with F(ab’), goat anti-human IgM antibodies (Jackson) to capture IgM-RNA complexes formed by the presence ‘of monoclonal anti-Ro(SS-A) antibodies. Results
Hybridoma fusion results Three tonsillar lymphocyte fusions, BOU (male, 7 years old), LIL (female, 8 years old) and LAP (male, 8 years old) were separately undertaken. The percent hybridoma growth varied with each tonsil fusion but ranged from 11 to 480/b and the percent of hybridoma IgM producers ranged from 55 to 69%. The remaining hybridomas produced IgG of unknown specificity or no immunoglobulin. These yields are comparable to published reports on successful human-human hybridoma fusions without the use of Epstein-Barr virus transformation [32]. Autoreactivity of hybridoma supernatants Undiluted hybridoma supernatants were initially screened by ELISA for the presence of autoantibodies against DNA, ssDNA, cardiolipin and Ro(SS-A). GM 4672 supernatant showed no autoreactivity and no reactivity was seen in wells lacking antigen or those containing normal human IgM or IgG standards or HGM. Overall, 8-147; @O/736) of the hybridomas showed autoreactivity with the screening antigens. Sixty-six percent (53/80) of the hybridomas from the three fusions reacted to only one of the three antigens used in the initial screening: 26% (21/80) bound ssDNA, 21°; (17/80) bound Ro(SS-A), 696 (5/80) bound cardiolipin and 13”,, (10/80) showed reactivity to both native and ssDNA. The remaining 34”; (27/80) showed reactivity to more than one of the three antigens, with 22/27 binding to Ro( SS-A). There was no correlation between IgM concentration (ranging from 0.31 to 20 ug/ml) and the presence of polyreactivity among these hybridomas. None of the initial hybridomas was screened for anti-La(SS-B) or anti-Sm/RNP binding. Autoreactivity of cloned supernatants Thirty of 39 hybridomas with anti-Ro( SS-A) reactivity were selected for cloning by limiting dilution. Successful cloning was observed in 21/30 cell lines, of which 13 retained their reactivity to Ro(SS-A). The calculated probability of the obtained cloned anti-Ro( SS-A) antibodies originating from a single cell was greater than 95 0 (, [331. The undiluted supernatants of these cloned anti-Ro( SS-A) positive antibodies were tested for binding to ssDNA, RNA, cardiolipin, BSA, caseine, IgG (whole, Fc and Fab’), La(SS-B) and Sm/RNP. None of these autoantibody supernatants showed reactivity to RNA, BSA or caseine. Five of 19 cloned anti-Ro(SS-A) monoclonal antibodies bound strongly in ELISA to La(SS-B) and/or to Sm/RNP (Figure 1). One, LAP 149.2, showed strong reactivity to ssDNA and another, BOU 11.4, to cardiolipin. Twelve showed reactivity to IgG. None reacted ,with IgG Fc only, and when strong binding was observed to whole IgG, binding to IgG Fab’ was equal to or greater than binding to IgG Fc. The IgM concentration of anti-Ro(SS-A)
Fob
1.0 Sm/RNP
LO
2.0 -
-1
1.0-
Ro
2.0
-
Card
:I; ssONA
Figure 1. Relative ELISA binding (OD,,, ) of cloned anti-Ro(SS-A) antibodies ( W LAP, 0 BOU, q LIL fusions) to ssDNA, cardiolipin (cardio), Ro(SS-A) (Ro), La(SS-B) (La), Sm/RNP, IgG and its Fc and Fab fragments. Also shown are the IgM concentration and light chain expression of these clones. All supematants were tested undiluted.
Normal human monoclonal anti-Ro( SS-A) antibody
777
reactive clones varied from 0.1 to 44.9 pg/ml (mean 5.09 ug/ml). Either K (1 l/ 19) or h light (8/19) chains were expressed on the IgM cloned antibodies (Figure 1). These anti-Ro( SS-A) antibody-producing clones were stable for more than 2 years before being frozen for later use.
ELISA
inhibition
Inhibition of Ro(SS-A) binding was performed on all cloned IgM anti-Ro(SS-A) antibodies to assess their specificity. Since the results were similar for all clones, the inhibition results of only four clones are presented. These clones predominantly reacted to Ro( SS-A) since diluted supernatants lost their reactivity to La( SS-B) and Sm/RNP while maintaining relatively strong anti-Ro(SS-A) binding. Only antiRo( SS-A) reactivity remained in the dilutions selected for inhibition assays. Representative inhibition curves with cloned supernatants or with SLE serum IgM or IgG anti-Ro(SS-A) antibodies are shown in Figures 2 and 3, respectively. No inhibition of IgM monoclonal or serum polyclonal IgM or IgG anti-Ro(SS-A) binding was observed with RNA, IgG, ssDNA, DNA or BSA. Among the clones represented here (Figure 2), inhibition of anti-Ro(SS-A) binding was observed with Ro(SS-A) itself in a dose-dependent manner and ranged from 74 to 88Og inhibition with 25 ug of Ro(SS-A). Similar dose-dependent inhibition of anti-Ro(SS-A) binding to Ro(SS-A) was also achieved with 25 ug of La(SS-B) and/or Sm/RNP. Inhibition of polyclonal serum IgM anti-Ro( SS-A) binding of BRA andASH with Ro(SS-A) was, respectively, 70:/, and460/,, with La(SS-B) 57”” and 520/b, and with Sm/RNP 527,” and 28?, at 25 ug of antigen (Figure 3). The inhibition of the same two patients’ serum polyclonal IgG anti-Ro(SS-A) binding was achieved with lower with 5 pg), while no inhibition was concentrations of Ro(SS-A) (100 9/, inhibition observed with up to 25 ug of La( SS-B) or Sm/RNP.
Immunoblotting The Coomassie blue- and silver-stained SDS-PAGE as well as the Amido blackstained nitrocellulose showed that the Ro(SS-A) preparation contained a single band of 60 kDa. The immunoblotting (Figure 4) included as negative controls, HGM (lane l), a normal human IgM serum (lane 15) and an ELISA anti Ro( SS-A) negative supernatant BOU 11.4 (lane 2). All the anti Ro(SS-A) ELISA positive clones as well as an anti-Ro(SS-A) ELISA IgM positive serum TOW (lane 14), as a positive control, and two anti-Ro(SS-A) ELISA positive hybridomas (lanes 12 and 13) showed variable intensity of binding to the 60 kDa band. The same blot, developed with an anti-IgG alkaline phosphatase conjugate, revealed, as expected, binding to the 60 kDa band by the TOW serum only (data not shown). The La(SS-B) preparation displayed three bands: two major bands at 42 and 14 kDa and a minor band at 29 kDa. All but three clones (BOU 77.1, LAP 234.8 and 177.4), which showed anti-La(SS-B) ELISA reactivity, showed weak binding to one or all three bands of immunoblotted La(SS-B) (data not shown). Such binding of anti-La(SS-B) antibodies to La(SS-B) degradation products (29 and 14 kDa) has been noted by other investigators [34-361. Only the SLE sera showed binding to La(SS-B) when the blot was developed with an anti-IgG conjugate.
778
Helene Massicotte et al. LAP
156.4
LIL
loo’- l
Qr
16.13
100
70-m \ 90'
90
80 -
SO 70
60
50’
40
30. 20 IO t
”
IPg BOU
5Pa
25 PP
53.4
100 go-
90 -
60-
60-
70-
70 -
60 -
60-
50-
50-
40-
40‘
30-
30 -
20-
20 IO -
IO -
0’
’
1 #P
I
5 ILP
I
255~
Figure 2. Inhibition of cloned IgM anti-Ro(SS-A) antibody binding. One, 5 and 25 pg of the following inhibitors were added: Ro(SS-A) (0), La(SS-B) (A), Sm/RNP ( q), RNA (0), ssDNA ( l ), DNA (A), IgG ( n ) and BSA (4). The IgM concentration (pg of IgM/ml) and the dilution factor of each cloned supematant chosen to obtain an optical density of 1.0 were for BOU 53.4 (0.7 pg/ml) l/SO, BOU 11.4 (1.63 pg/ml) l/5, LAP 158.4 (2.96pg/ml) l/4 and LIL 16.13 (1.5 pg/ml) l/4.
None of the anti-Ro(SS-A) positive clones that bound to Sm/RNP by ELISA blotted any proteins associated with the Sm/RNP preparation (68kDa, A, BB’, C, D, E, F and G). Positive controls in this immunoblotting experiment consisted of clones selected specifically for anti-Sm/RNP binding and sera with anti-Sm/RNP reactivity (data not shown). The staining procedures used showed no cross-contaminating protein bands of Ro(SS-A), La(SS-B) or Sm/RNP by each other, or by IgG.
Normal human mon&lo&l BRA-IqM
._oI z
1u9
z m
8
anti-Ro(SS-A)
antibody
779
BRA-IqG t
IO*
5u9
25P9
0.1 ug
ASH-IqM
5P9
15co
ASH-IgG iooy-
IOOgoBO -
60504030 20 IO t
01
’
I
I
1 UP
5Pq
25~~
Figure 3. Inhibition of IgM anti-Ro(SS-A) binding of sera (dilution) BRA (l/80) and ASH (1 /SO) and of IgG anti-Ro(SS-A) binding of the same sera (l/10,000). The following inhibitors were added: Ro(SS-A) (Z), La(SS-B) (a), Sm/RNP (U), RNA (O), ssDNA (O), DNA (A), IgG (m) and BSA (*). The dilution of each serum was chosen to obtain an optical density of 1 .O.
Only those anti-Ro(SS-A) clones that bound IgG, Fab’ and/or Fc by ELISA showed binding to IgG components in immunoblotting experiments. These antibodies bound to both the heavy and light chain of IgG as well as to Fc and Fab’ fragments (data not shown). In immunoblots of normal lymphocyte lysates, only one monoclonal antibody, BOU 43.10, showed weak binding to bands of 79, 61, 42 and 13 kDa, which could correspond to Ro( S S-A) and La( S S-B). This could not be confirmed by inhibition studies because of the weak reactivity (data not shown). The same bands were also
780 Helene Massicotte et al.
Figure 4. Immunoblotting Ro(SS-A) with supernatants and sera. The amount of purified Ro(SS-A) (J. B. Harley) used per lane was 5ug. The undiluted cloned supematants (IgM concentration) or serum (dilution) used in each lane were as follows: 1: HGM, 2: BOU 11.4 (0.12 ug/ml), 3: BOU 53.6 (21.08 ug/ml), 4: BOU 77.1 (6.25 ug/ml), 5: LAP 234.5 (5.95 ug/ml) 6: BOU 318.1 (1.33 ug/ml), 7: LAP 177.4 (1.5 ug/ml), 8: LAP 193.1 (2.06 ug/ml) 9: LAP 149.2 (52.66 ug/ml) 10: LIL 13.12 (0.35 ug/ml), 11: BOU43.10 (6.35 ug/ml), 12: LIL 77 (3.28 ug/ml), 13: BOU 301 (5.84 ug/ml) 14: TOW (l/SO) 15: Normal serum (l/50). The F(ab’), goat anti-human IgM alkaline phosphatase conjugate (Jackson) and BCIP/NBT substrate (Gibco) were used to detect bound IgM.
seen when SLE sera with anti-Ro(SS-A) and anti-La(SS-B) reactivity were used. None of the other monoclonal antibodies showed binding to any bands in these cell lysates.
Immunoprecipitation Ten monoclonal IgM anti-Ro(SS-A) antibodies were tested for reactivity to the hYRNAs by immunoprecipitation with HeLa cell lysates. None of them precipitated hYRNAs. LIL 72.52 precipitated tRNA, 4.5s and 5s RNAs and a unique unidentified high molecular weight RNA; LIL 16.13 precipitated Ul and U2 RNAs and 7s RNA, consistent with its Sm/RNP ELISA reactivity (data not shown).
Discussion The present study shows that antibodies reactive with Ro(SS-A) can be successfully obtained by the technique of somatic cell hybridization. Anti-Ro(SS-A) antibodies
Normal human monoclonal anti-Ro(SS-A)
antibody
781
were seen in each of the three separate fusions performed (thus reproducible) employing lymphocytes from tonsils of three different normal individuals. This observation implies the existence of immunoglobulin genes encoding anti-Ro( SS-A) autoantibodies in normal humans. The present study further extends earlier observations from this laboratory on the generation of normal human:human hybridomas with autoreactivity [26]. This is also in keeping with the observations of others, that normals have the potential to produce natural antibodies to a variety of self antigens [37-40]. The binding properties of the normally-derived monoclonal IgM antiRo(SS-A) antibodies appear similar to the polyclonal serum IgM anti-Ro(SS-A) antibodies of patients with SLE. Monoclonal anti-Ro(SS-A) antibodies were selected by ELI SA for their reactivity to Ro(SS-A). Some of these monoclonal antibodies also reacted strongly with La(SS-B) or Sm/RNP, but not with RNA, BSA or caseine. However, the binding reactivity to La( SS-B) and Sm/RNP was weaker than binding to Ro(SS-A), since the former reactivity could be readily diluted out. Ro(SS-A), La(SS-B) and Sm/RNP were nonetheless all able to inhibit the Ro(SS-A) binding of these monoclonal anti-Ro(SS-A) antibodies in a dose-dependent manner at dilutions of antibody supernatant in which direct ELISA binding of La(SS-B) and Sm/RNP were undetectable. All of the monoclonal anti-Ro( SS-A) antibodies recognized a structural epitope on the 60 kDa Ro(SS-A) protein, and many also recognized a structural epitope on the polypeptides of the La(SS-B) protein. A presumed conformational epitope on Sm/RNI? proteins was recognized by some of the monoclonal antiRo(SS-A) antibodies, since those that reacted with Sm/RNP in ELISA failed to immunoblot Sm/RNP proteins. This may be explained by the observation that the intensity of binding reactivity to any antigen by ELISA or Western blotting does not always correlate [ 10,131, because these assays present the antigen in different forms to the antibodies. Thus, ELISA presents antigen as a native molecule in either a bound or free aqueous (inhibition) phase while Western blotting presents antigen in a linear form (denatured and reduced). None of these monoclonal anti-Ro(SS-A) antibodies immunoprecipitated hYRNAs, although one clone, which also bound Sm/RNP by ELISA, immunoprecipitated Ul and U2 RNA from HeLa cells. These results could be a reflection of the lack of sensitivity of the immunoprecipitation assay with HeLa cells by the relatively low concentration and affinity of the monoclonal IgM anti-Ro(SS-A) antibodies or because of lack of access to the epitope on the whole RNA-protein complex recognized by these particular monoclonal anti-Ro(SS-A) antibodies. Similar dose-dependent inhibition of Ro(SS-A) binding by Ro(SS-A), La(SS-B) and Sm/RNP was observed with the two representative polyclonal SLE serum IgM anti-Ro(SS-A) antibodies. In contrast, SLE serum IgG anti-Ro(SS-A) antibody binding was specifically inhibited with Ro(SS-A) and not with La(SS-B) or SmlRNP. These findings indicate that the property of polyreactivity is shared by patient serum IgM antibody and the normally-derived monoclonal IgM antibodies. These monoclonal anti-Ro( SS-A) antibodies also reacted to whole IgG or to its Fc and Fab fragments. This reactivity was clearly directed towards structural epitopes of the IgG molecule as revealed by the ability of these monoclonal IgM antiRo( SS-A) antibodies to blot both the heavy and light chains of IgG as well as its Fc and Fab’ fragments. However, since IgG, Fc or Fab’ could inhibit the Ro(SS-A)
782
Helene
Massicotte
et al.
binding, different sub-regions of the H and/or L chain may be involved in the reactivity of these antibodies to Ro( SS-A) and IgG. The binding to Ro( SS-A) could involve, for example, the VH chain region, while the rheumatoid factor activity, as recently described by Weisbart et al. [41], could involve the VL region chain. Particular H and L chain pairings can also alter the binding of the recombinant antibodies compared with the binding properties anticipated from the individual H and L chains [42]. Goldman et al. [43] described monoclonal rheumatoid factors that react with both the Fab and Fc region of autologous IgG. Anti-Fab antibodies have also been frequently noted amongst patients with rheumatoid arthritis [44] and in some instances closely associated with the presence of antibody reactivity to Ro(SS-A) [45]. Indeed, Ro(SS-A) and IgG Fab appear to share epitopes that are bound by polyclonal serum IgG anti-Ro(SS-A) antibody [45]. Further, it has been shown in mice that monoclonal antibodies selected for rheumatoid factor activity have polyspecificity to other unrelated autoantigens [46]. In a previous study from this laboratory, normal human monoclonal IgM antibodies selected for DNA binding frequently revealed polyreactivity to a variety of nucleic acid antigens, including DNA, cardiolipin and synthetic polynucleotides with different base structures [26], but none showed binding to the RNA-protein associated antigens (Ro(SS-A), La(SS-B) and Sm/RNP; data not shown). Similarly, the antibodies selected here for their reactivity to Ro(SS-A) also showed polyreactive binding to other unrelated RNA-protein associated antigens, but rarely to DNA related antigens (2/19 anti-Ro(SS-A) antibodies bound DNA) and uncommonly to cardiolipin (7/19). This dichotomy parallels the observations in patients with where anti-RNA-protein-associated IgG polyclonal autoantibody response, autoantibodies and anti-DNA autoantibodies often appear independently. The present study does not explain why these monoclonal IgM antibodies bound to what appear to be structurally unrelated RNA-protein complexes. There is no known shared structural epitope on Ro( SS-A) and La( SS-B) recognized at present, however, Ro(SS-A) and La(SS-B) are often found as a complex [20, 21, 23, 241. Moreover, a significant proportion of patients have serum IgG antibodies that bind to Ro(SS-A) as well as to La(SS-B) [2, 11, 12,20-241. It is conceivable, therefore, that the normal immune system may have B cells whose Ig receptor recognizes Ro(SS-A) and La(SS-B) individually, as a complex, or in some instances may recognize some epitope common to both of these molecules. In summary, the present study reveals that normal human B cells have the genetic potential to produce antibodies to Ro(SS-A). Some monocional anti-Ro(SS-A) antibodies also strongly bound by ELISA to La( SS-B), Sm/RNP, IgG, Fab’ and Fc. Immunoblotting and inhibition studies disclosed reactivity of the same IgM molecules to both Ro(SS-A) and La(SS-B) proteins. This suggests the possibility that there is either a common shared protein epitope on the Ro(SS-A) and La(SS-B) antigens or that different subsites exist for the Ro(SS-A) and La(SS-B) antigens within the same antibody variable region. In contrast, the reactivity of these antibodies to Sm/RNP by ELISA may reflect binding to conformational determinants, since these antibodies did not reveal binding to Sm/RNP proteins by immunoblotting. It remains to be determined whether the normally-derived IgM monoclonal anti-Ro(SS-A) autoantibodies and the polyclonal serum IgM anti-
Normal human monoclonal anti-Ro( SS-A) antibody 783 Ro(SS-A) autoantibodies from SLE patients, which have similar properties, are derived from and share similar germline immunoglobulin V region genes. It is also controversial whether or not natural antibodies, which appear in normal individuals, are derived from the same precursor B cells that give origin to the IgG antibodies which either are expressed in disease or have pathogenic potential [47, 481. An analysis and comparison of the variable regions of the normal IgM anti-Ro(SS-A) antibodies produced by these hybridoma B cells and the IgG anti-Ro(SS-A) autoantibodies produced in disease should provide some insight into the factors responsible for determining their relatedness, expression and regulation.
Acknowledgements Dr Massicotte’s work was supported by a postdoctoral Fellowship from the Canadian Arthritis Society. This work is supported by a research grant from the Canadian Arthritis Society. Dr Harley is supported by the NIH and the Veterans Administration.
References 1. Lerner, M. R. and J. A. Steitz. 1979. Antibodies to small nuclear RNAs complexed with proteins are produced by patients with systemic lupus erythematosus. Proc. Natl. Acad. Sci. USA 76: 5495-5499 2. Harley, J. B., E. L. Alexander, W. B. Bias, 0. F. Fox, T. T. Provost, M. Reichlin, H. Yamagata, and F. C. Arnett. 1986. Anti-Ro/SS-A and anti-La/SS-B in patients with Sjogren’s Syndrome. Arthritis Rheum. 29: 196-206 3. Watson, R. M., A. T. Lane, N.K. Bamett, W. B. Bias, F. C. Amett, and T. T. Provost. 1984. Neonatal lupus erythematosus: a clinical, serological and immunogenetic study with review of the literature. Medicine (Baltimore) 63: 362-378 4. Lee, L. A., C. E. Harmon, J. C. Huff, D. A. Norris, and W. L. Weston. 1985. The demonstration of SS-A/Ro antigen in human fetal tissues and in neonatal and adult skin. J. Invest. Dermatol. 85: 143-146 5. Deng, J. S., R. D. Sontheimer, and J. N. Gilliam. 1985. Expression ofRo/SS-A antigen in human skin and heart.J. Invest. Dermatol. 85: 412-416 6. Harmon, C. H., J. S. Deng, C. L. Peebles, and M. E. Tan. 1984. The importance of tissue substrate in the SS-A/Ro antigen-antibody system. Arthritis Rheum. 27: 166-173 7. Wolin, S. L. and J. A. Steitz. 1983. Genes for two small cytoplasmic Ro RNAs are adjacent and appear to single copy in the human genome. Cell 32: 735-744 8. Kato, N., H. Hoshino, and F. Harada. 1982. Nucleotide sequence of 4.5s RNA (C8 or HY5) from HeLa cells. Biochem. Biophys. Res. Commun. 108: 363-370 9. Ben-Chetrit, E., E. K. L. Chan, K. F. Sullivan, and E. M. Tan. 1988. A 52 kd protein is a novel component of the SS-A/Ro antigenic particle. J. Exp. Med. 167: 1560-1571 10. Rader, M. D., C. O’Brien, Y. Liu, J. B. Harley, and M. Reichlin. 1989. Heterogeneity of the Ro/SS-A antigen: different molecular forms in lymphocytes and red blood cells. J. Gin. Invest. 83: 1293-1298 11. Woiin, S. L. and J. A. Steitz. 1984. The Ro small cytoplasmic ribonucleoprotein: identification of the antigenic protein and its binding site on the Ro RNAs. Proc. Nacl. Acad. Sci. USA 81: 1996-2000 12. Lieu, T. S., M. Jiang, J. C. Steigerwald, and E. M. Tan. 1984. Identification of the SS-A/Ro intracellular antigen with autoimmune sera.g. Zmmunol. Methods 71: 217-228 13. Yamagata, H., J. B. Harley, and M. Reichlin. 1984. Molecular properties of the Ro/SS-A antigen and enzyme-linked immunosorbent assay for quantitation of antibody. J. Clin. Invest. 74: 625-633
784
Helene
Massicotte
et al.
14. McNeilage, L. J., S. Whittingham, and I. R. MacKay. 1984. Autoantibodies reactive with small ribonucleoprotein antigens: a convergence of molecular biology and clinical immunology. 3. Clin. Lab. Zmmunol. 15: 1-17 15. Reddy, R., D. Henning, E. M. Tan, and H. Busch. 1983. Identification of a La protein binding site in a RNA polymerase III transcript (4.5 IRNA). 3. Biol. Chem. 258: 8352-8356 16. Madore, S. J., E. D. Wieben, and T. Pederson. 1984. Eukaryotic small ribonucleoproteins.3. Biol. Chem. 259: 1929-1933 17. Kurilla, M. G. and J. D. Keene. 1983. The leader RNA of vesicular stomatitis virus is bound by a cellular protein reactive with anti-La lupus antibodies. Cell 34: 837-845 18. Habets, W. J., J. H. den Brok, A. M. Th. Boerbooms, L. B. van de Putte, and W. J. van Venrooij. 1983. Characterization of the SS-B (La) antigen in adenovirus-infected and uninfected HeLa cells. EMBOJ. 2: 1625-1631 19. Francoeur, A. M., E. K. L. Chan, J. I. Garrels, and M. B. Mathews. 1985. Characterization and purification of lupus antigen La, an RNA binding protein. Mol. Cell. Biol. 5: 586-590 20. Eisenberg, R. A. 1985. Association between Ro and La antigenic determinants: immunodiffusion analysis of human spleen extract. 3. Zmmunol. 3 5: 1707- 17 13 21. Hardin, J. A. 1986. The lupus autoantigens and the pathogenesis of systemic lupus erythematosus. Arthritis Rheum. 29: 457-460 22. Boire, G. and J. Craft. 1989. Biochemical and immunological heterogeneity of the Ro ribonucleoprotein particles. Analysis with sera specific for Rohy5 particle. 3. Clin. Invest. 84: 270-279 23. Boire, G. and J. Craft. 1990. Human Ro ribonucleoprotein particles: characterization of native structure and stable association with the La polypeptide. 3. Clin. Invest. 85: 1182-1190 24. Mamula, M. J., E. D. Silverman, R. M. Laxer, L. Bentur, B. Isacovics, and J. A. Hardin. 1989. Human monoclonal anti-La antibodies. The La protein resides on a subset of Ro particles. 3. Zmmunol. 143: 2923-2928 25. Cairns, E., J. St Germain, and D. A. Bell. 1985. The in vitro production of anti-DNA antibody by cultured peripheral blood or tonsillar lymphoid cells from normal donors and SLE patients. 3. Immunol. 13: 3839-3844 autoantibody-producing 26. Cairns, E., J. Block, and D. A. Bell. 1984. Anti-DNA hybridomas of normal human lymphoid cell origin. 3. C/in. Invest. 74: 880-887 27. Gaither, K. K., 0. F. Fox, H. Yamagata, M. J. Mamula, M. Reichlin, and J. B. Harley. 1987. Implications of anti-Ro/Sjogren’s Syndrome A antigen autoantibody in normal sera for autoimmunity. 3. Clin. Invest. 79: 841-846 28. Harley, J. B., H. Yamagata, and M. Reichlin. 1984. Anti-La/SS-B antibody is present in some normal sera and is coincident with anti-Ro/SSA precipitins in systemic lupus erythematosus.3. Rheum. 11: 309-314 29. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the bacteriophage T4. Nature 227: 680-685 30. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoresis transfer of protein from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76: 4350-4360 31. Forman, M. S., M. Nakamura, T. Mimori, C. Gelpi, and J. A. Hardin. 1985. Detection of antibodies to small nuclear ribonucleoproteins and small cytoplasmic ribonucleoproteins using unlabeled cell extracts. Arthritis Rheum. 28: 1356-I 361 32. Massicotte, H., J. Rauch, Y. Shoenfeld, and H. Tannenbaum. 1984. Influence of fusion cell ratio and cell plating density on production of human-human hybridomas secreting anti-DNA autoantibodies from patients with systemic lupus erythematosus. Hybridoma 3: 215-222 33. Coller, H. A. and B. S. Coller. 1983. Statistical analysis of repetitive subcloning by the limiting dilution technique with a view toward ensuring hybridoma monoclonality. Hybridoma 2: 91-96
Normal human monoclonal anti-Ro( SS-A) antibody
785
of 34. Venables, P. J. W., P. R. Smith, and N. Maini. 1985. Purification and characterization the Sjogren’s Syndrome A and B antigens. Clin. Exp. Immunol. 5: 731-738 35. Harley, J. B., M. 0. Rosario, H. Yamagata, 0. F. Fox, and E. Koren. 1985. Immunologic and structural studies of the lupus/Sjogren’s Syndrome autoantigen, La/SSB, with a monoclonal antibody. J. Clin. Znvesr. 76: 801-806 and E. M. Tan. 1986. Epitopes, structural domains, 36. Chan, E. K. L., A. M. Francoeur, and assymmetry of amino acid residues in SS-B/La nuclear protein. 3. Zmmunol. 136: 3744-3749 B. G., G. Dighiero, and S. Avrameas. 1983. Naturally occuring antibodies 37. Guilbert, against nine common antigens in human sera. 3. Immunol. 128: 2779-2787 R. G. Whalen, 38. Dighiero, G., P. Lymberi, J.-C. Mazie, S. Rouyre, G. S. Butler-Browne, and S. Avrameas. 1983. Murine hybridomas secreting natural monoclonal antibodies reacting with self antigens. 3. Immunol. 131: 2267-2272 the “know thyself’ of 39. Avrameas, S. 1986. Natural autoreactive B cells and autoantibodies: the immune system. Ann. Zmmunol. Inst. Pasteur 137D: 150-156 D. M., R. A. Eisenberg, and A. D. Steinberg. 1990. Development of the 40. Klinman, autoimmune B cell repertoire in MRL-lpr/lpr mice. 3. Zmmunol. 144: 506-571 41. Weisbart, R. H., A. L. Wong, D. Noritake, A. Kacena, G. Chan, C. Ruland, E. Chin, I. S. Y. Chen, and J. D. Rosenblatt. 1991. The rheumatoid factor reactivity of a human IgG monoclonal autoantibody is encoded by a variant VkII L chain gene.3. Zmmunol. 147: 2795-2801 42. Radic, M. Z., M. A. Mascelli, J. Erikson, H. Shan, and M. Weigert. 1991. Ig H and L chain contributions to autoimmune specificities.3. Zmmunol. 146: 176-182 M., J. C. Renversez, and I’. H. Lambert. 1983. Pathological expression of 43. Goldman, idiotypic interactions: immune complexes and cryoglobulins. Springer Semin. Immunopathol. 6: 33-49 of 44. Nasu, H., D. Chia, Y. Iwaki, E. V. Barnett, and P. I. Terasaki. 1981. Cytotoxicity anti-Fab antibodies against B lymphocytes in rheumatoid arthritis. Arthritis Rheum. 24: 1278-1282 45. Mamula, M. J., 0. F. Fox, H. Yamagata, and J. B. Harley. 1986. The Ro/SS-A autoantigen as an immunogen: some anti-Ro/SS-A antibody binds IgG. 3. Exp. Med. 86: 1889-1901 46. Rubin, R. L., R. S. Balderas, E. M. Tan, F. J. Dixon, and A. N. Theofilopous. 1984. Multiple autoantigen binding capabilities of mouse monoclonal antibodies selected for rheumatoid factor activity. 3. Exp. Med. 159: 1429-1440 1987. The genetic origin of 47. Kofler, R., F. J. Dixon, and A. N. Theofilopoulos. autoantibodies. Immunol. Today 8: 374-48 48. Bona, C. A. 1988. V genes encoding autoantibodies: molecular and phenotypic characteristics. Annu. Rev. Immunol. 6: 327-358
Characterization of Human-Human Hybridoma Monoclonal Anti-Ro( SS-A) Autoantibodies Derived from Normal Tonsil Lymphoid Cells
Helene Massicotte, John B. Harley* and David A. Bell University University
of Western Ontario, Rheumatic Hospital,
Program, University
London,
Oklahoma
Ontario,
Medical
Disease Unit, Department
Canada and *The Arthritis
Research
Foundation,
Department
of Medicine,
and Immunology of Medicine,
of Oklahoma Health Sciences Center and Veterans Administration Medical Center, Oklahoma City, OK, USA (Received 12 February 1992 and accepted 3 August 1992)
Human-human hybridomas obtained from the separate fusion of tonsillar lymphoid cells from three different normal individuals to the lymphoblastoid cell line GM 4672 were screened by ELISA for the presence of autoantibody to Ro(SS-A). Those anti-Ro(SS-A) reactive hybridomas were then cloned by limiting dilution. Nineteen monoclonal IgM antiRo(SS-A) antibodies were obtained, which showed specificity to Ro(SS-A) by ELISA and Western blotting (60 Mla). Some of these monoclonal antiRo(SS-A) antibodies showed reactivity to DNA (2/19), cardiolipin (9/19), Sm/RNP (15/19) by ELISA, and to IgG (12/19) and La(SS-B) (19/19) by ELISA and Western blotting. None showed reactivity to the unrelated proteins caseine and BSA, nor to RNA. Inhibition studies revealed that the binding to Ro( SS-A) of both IgM hybridoma monoclonal and SLE serum polyclonal IgM anti-Ro(SS-A) antibodies was inhibited with Ro(SS-A), La( SS-B) and Sm/RNP but not with IgG, DNA, RNA and BSA. These data indicate that (1) normal humans have the genetic potential to express antibodies to Ro(SS-A) and (2) the normally derived monoclonal and SLE serum IgM anti-Ro(SS-A) antibodies share similar antigen binding properties and therefore may possibly originate from a common pool of precursor B cells.
Introduction
Systemic lupus erythematosus (SLE) is characterized diverse autoantibodies. Among these, autoantibodies
by the presence of many to nucleic acid antigens
Correspondence to: Dr David A. Bell, University Hospital, Rheumatic Disease Unit, Box 5339, London, Ontario, Canada, N6A 5A5. 771 0896-841 l/92/060771
+ 15 $08.00/O
0 1992 Academic Press Limited
772 Helene Massicotte et al. (DNA/histone complex) have been well studied because of their observed relationship to and association with, tissue injury and disease activity. More recently, increasing interest has focused on a variety of other autoantibodies directed against RNA-protein complexes e.g. Ro(SS-A), La(SS-B), Sm, UlRNP, whose presence identifies certain subsets of disease and/or related connective tissue syndromes such as SLE [l] and Sjogren’s syndrome [2] or occurs in mothers of infants with neonatal SLE or congenital heart block [3]. The investigation of these latter autoantibodies has accelerated recently with the development of greater insights into the molecular nature of the target autoantigens of these RNA-protein complexes. Briefly, it has been demonstrated that Ro(SS-A), a constituent of many normal tissue cells [P61, is present on a particle comprised of RNA of 83-112 bases (hyl, hy2, hy4 and hy5 in man) [7,8] and one or more associated polypeptides (60,54 and/or 52 kDa, depending on the cell type) [9, lo] which bear the antigenic epitopes [ 1 l-l 31. La( SS-B), another eukaryotic small RNP particle, contains RNA of about 80-l 30 nucleotides including precursors of RNA polymerase III transcripts [ 14,151, Ul RNA [ 161, certain small viral RNAs [ 171 and a highly conserved 43-50 kDa phosphoprotein [18, 191. Ro(SS-A) may exist alone or complexed to La(SS-B) depending on the YRNA present [2,11-12,20-241. The ability spontaneously to generate human hybridoma monoclonal autoantibodies to other nuclear antigens by the fusion of tonsillar lymphocytes of a normal individual to the human lymphoblastoid cell line GM 4672 [25,26], and the observation of low levels of anti-Ro(SS-A) antibodies in normal donor plasma [27], raised the possibility that natural antibodies to Ro(SS-A) could be produced in vitro. The present paper shows that autoantibodies to Ro(SS-A) can be successfully obtained by this strategy of somatic cell hybridization with normal human lymphocytes and explores the reactivity and specificity of these anti-Ro(SS-A) antibodies with other RNA-protein complexes, DNA, IgG and cardiolipin.
Materials and methods 1. Production of human :human hybridomas The production and propagation of humanhuman hybridoma autoantibodies using normal whole tonsillar lymphoid cells fused to the human lymphoblastoid cell line GM 4672 was performed as previously described [26]. The tonsils used were from normal children who did not demonstrate anti-Ro(SS-A) antibodies in their serum and had no family history of connective tissue disorder.
2. ELISA
autoantibody
screening
Both IgM and IgG autoantibody production were screened by ELISA in undiluted supernatants of all hybridomas and their clones, using alkaline phosphatase conjugates, F(ab’), goat anti-human IgM (u chain specific) and F(ab’), goat anti-human IgG (y chain specific) respectively (Jackson Immunoresearch Laboratory, Mississauga, Ont, Canada). The ELISA used for the detection of antibodies to ssDNA, DNA, RNA and cardiolipin were described previously [25,26]. Standard sera of KAPPS were used as controls for the cardiolipin assay.
Normal human monoclonal anti-Ro(SS-A)
antibody
773
The affinity-purified calf thymus or spleen Ro(SS-A) [2, 13,27,28], the afBnitypurified bovine thymus or spleen La(SS-B) and Sm/RNP (Immunovision, Springdale, AR) were coated overnight at 4°C in carbonate coating buffer, pH 9.6 at 3 pg/ml. The supernatants as well as the conjugate were both subsequently incubated on the plates overnight at 4°C. The washing buffer contained O.lq/, bovine serum albumin (BSA) (Boehringer Mannheim, Gmbh, Germany) in 0.1 M Tris-HCl, 0.059b Tween 20, pH 7.4 (Fisher Scientific Co, Fair Lawn, NJ, USA). The binding of hybridoma IgM antibodies to human IgG (rheumatoid factor ELISA assay) used whole human IgG (Cappel, Organon Teknika Inc, Scarborough, Ont., Canada), Fc or Fab’ fragments of IgG (Jackson) coated at 10 pg/ml in carbonate coating buffer. A diluted (in 1 o/0bovine serum albumin (BSA)-saline) rheumatoid factor-positive serum control or undiluted supernatants were added to antigen coated plates, previously blocked with 3% BSA-saline, and incubated overnight at 4°C. F(ab’), goat anti-IgM alkaline phosphatase conjugate (Jackson) was used at a dilution of 1:2,500 in 30/6BSA-0.05% Tween-salineandleft for 3 hat roomtemperature (RT). All anti-Ro(SS-A) antibody cloned supernatants were also tested for reactivity to the unrelated antigens BSA and caseine (BDH Chemicals Ltd, Poole, UK), coated at 40 ug/ml in carbonate coating buffer or to wells without antigen. 3. Sera SLE sera (5) with IgG anti-Ro(SS-A) antibody reactivity were studied in order to compare their IgM and IgG ELISA binding to Ro( SS-A), La( SS-B) and Sm/RNP, and to perform inhibition assays. The results of two representative sera, BRA and ASH, are shown in Figure 3. ASH serum by ELISA had IgM and IgG antibodies to Ro(SS-A), La(SS-B) and Sm/RNP; BRA serum had IgM antibodies to all three antigens and IgG antibodies to Ro(SS-A) and La(SS-B) only. 4. Inhibition of ELISA
Ro (SS-A)
binding
Inhibition assays were performed to confirm the Ro(SS-A) specificity of the cloned antibodies. Anti-Ro(SS-A) antibodies from sera or from cloned supernatants were premixed in a test tube with either phosphate buffered saline (PBS; 100% binding value), Ro(SS-A), La(SS-B), Sm/RNP, RNA, BSA, ssDNA, DNA, or whole IgG at 25, 5 and 1 gg for IgM antibody inhibition or 15,5, 1 and 0.1 gg for IgG antibody inhibition. The mixture was incubated overnight at 4°C then plated on Ro(SS-A)coated plates previously blocked with hybridoma growth medium (HGM) for 2 h at RT. The mixture was left on the plates for 3 h at 4°C then washed off and the conjugate added as previously described. The percent binding was calculated as follows: [(OD405llln+ inhibitor)/(OD),,nm + PBS)] x 100. 5. Heavy and light chain class expression of hybridoma monoclonal antibodies To determine and quantify the heavy chain (l.tor y chain) expressed in the hybridoma supernatants, either F(ab’), goat anti-human IgM or IgG (heavy chain-specific, Jackson) were coated in 10 pg/ml carbonate coating buffer. Serially diluted
774
Helene Massicotte et al.
supernatants as well as IgM and IgG standards (Cappel, range 0.01 to 2 p,g/ml) were incubated on the coated plates overnight at 4°C. The conjugates used were, respectively, a goat anti-human IgM (Zymed Laboratories, Mississauga, Ont., Canada) or goat anti-human IgG (Kirkergaard Laboratories, Mandel Scientific Company Ltd, Guelph, Ont., Canada) alkaline phosphatase diluted 1:2,000 in 0.1 y0 BSA, 0.1 M Tris-HCl pH 7.4 and incubated for 3 h at RT. The curves established with the standards were used to determine the concentrations of IgM or IgG in the hybridoma supernatants. The light chains expressed by and associated with the IgM antibodies produced in the hybridoma supernatants were determined in a similar fashion. F(ab’), goat anti-human IgM (Jackson) was coated as above. A goat anti-human lambda (h) or kappa (K) chain-specific alkaline phosphatase conjugate (Sigma) diluted at 1:6,000 in 1 Y0 BSA, 2% bovine gamma globulin (BGG), 0.1 M Tris-HCl pH 7.4 (BGG, Cohn fraction II, III, Sigma) was used to develop this assay. The GM 4672 supernatant (IgG K) was used as a negative control and two Waldenstrom sera (IgM h or IgM K) diluted 1: 1,000 were employed as positive controls for h and K light chain assays, respectively. 6. SDS-PAGE Ro(SS-A), La(SS-B), Sm/RNP, IgG (whole, Fc and Fab’) and normal human lymphocyte lysates were separated by SDS-PAGE (sodium dodecyl sulfatepolyacrylamide gel electrophoresis). These antigens were reduced by boiling for 2 min at 96°C in 5% p-2 mercaptoethanol (Sigma), and electrophoresed according to Laemmli [29] with 15% and 5% polyacrylamide separating and stacking gels, respectively, both with 0.1% SDS. The SDS-PAGE was performed in a MiniProtean II apparatus (BioRad Chemical Division, Mississauga, Ont., Canada) according to the manufacturer. Prestained molecular weight markers (Gibco/BRL, Burlington, Ont., Canada) were run on the gel along with the antigens. 7. Immunoblots
The proteins separated by SDS-PAGE were then transferred electrophoretically from the gel to nitrocellulose (BioRad) [30] in a Mini-Trans Blot cell apparatus (BioRad) according to the manufacturer, at 100 V for 1 h with stirring and cooling. The nitrocellulose cut into strips was blocked overnight at 4°C with 5% powdered skim milk in Tris-buffered saline (TBS). Next, the strips were washed with TBS and incubated with undiluted cloned supernatants or control positive SLE or normal sera for 3 h at RT. After extensive washing, the F(ab’), goat anti-human IgM alkaline phosphatase conjugate (Jackson) at 1:2,500 in 1 96 BSA-TBS was added for a further incubation of 2 h at RT. The strips were again washed and then developed with NBT/BCIP substrate (Gibco/BRL) and the reaction stopped by water washes, and read. 8. Immunoprecipitation The immunoprecipitation, electrophoresis and silver staining of the RNAs immunoprecipitated by 10 of the monoclonal anti-Ro(SS-A) antibodies was performed according to Forman [31] with the following modification. Protein A beads were
Normal human monoclonal anti-Ro(SS-A)
antibody
775
precoated with F(ab’), goat anti-human IgM antibodies (Jackson) to capture IgM-RNA complexes formed by the presence ‘of monoclonal anti-Ro(SS-A) antibodies. Results
Hybridoma fusion results Three tonsillar lymphocyte fusions, BOU (male, 7 years old), LIL (female, 8 years old) and LAP (male, 8 years old) were separately undertaken. The percent hybridoma growth varied with each tonsil fusion but ranged from 11 to 480/b and the percent of hybridoma IgM producers ranged from 55 to 69%. The remaining hybridomas produced IgG of unknown specificity or no immunoglobulin. These yields are comparable to published reports on successful human-human hybridoma fusions without the use of Epstein-Barr virus transformation [32]. Autoreactivity of hybridoma supernatants Undiluted hybridoma supernatants were initially screened by ELISA for the presence of autoantibodies against DNA, ssDNA, cardiolipin and Ro(SS-A). GM 4672 supernatant showed no autoreactivity and no reactivity was seen in wells lacking antigen or those containing normal human IgM or IgG standards or HGM. Overall, 8-147; @O/736) of the hybridomas showed autoreactivity with the screening antigens. Sixty-six percent (53/80) of the hybridomas from the three fusions reacted to only one of the three antigens used in the initial screening: 26% (21/80) bound ssDNA, 21°; (17/80) bound Ro(SS-A), 696 (5/80) bound cardiolipin and 13”,, (10/80) showed reactivity to both native and ssDNA. The remaining 34”; (27/80) showed reactivity to more than one of the three antigens, with 22/27 binding to Ro( SS-A). There was no correlation between IgM concentration (ranging from 0.31 to 20 ug/ml) and the presence of polyreactivity among these hybridomas. None of the initial hybridomas was screened for anti-La(SS-B) or anti-Sm/RNP binding. Autoreactivity of cloned supernatants Thirty of 39 hybridomas with anti-Ro( SS-A) reactivity were selected for cloning by limiting dilution. Successful cloning was observed in 21/30 cell lines, of which 13 retained their reactivity to Ro(SS-A). The calculated probability of the obtained cloned anti-Ro( SS-A) antibodies originating from a single cell was greater than 95 0 (, [331. The undiluted supernatants of these cloned anti-Ro( SS-A) positive antibodies were tested for binding to ssDNA, RNA, cardiolipin, BSA, caseine, IgG (whole, Fc and Fab’), La(SS-B) and Sm/RNP. None of these autoantibody supernatants showed reactivity to RNA, BSA or caseine. Five of 19 cloned anti-Ro(SS-A) monoclonal antibodies bound strongly in ELISA to La(SS-B) and/or to Sm/RNP (Figure 1). One, LAP 149.2, showed strong reactivity to ssDNA and another, BOU 11.4, to cardiolipin. Twelve showed reactivity to IgG. None reacted ,with IgG Fc only, and when strong binding was observed to whole IgG, binding to IgG Fab’ was equal to or greater than binding to IgG Fc. The IgM concentration of anti-Ro(SS-A)
Fob
1.0 Sm/RNP
LO
2.0 -
-1
1.0-
Ro
2.0
-
Card
:I; ssONA
Figure 1. Relative ELISA binding (OD,,, ) of cloned anti-Ro(SS-A) antibodies ( W LAP, 0 BOU, q LIL fusions) to ssDNA, cardiolipin (cardio), Ro(SS-A) (Ro), La(SS-B) (La), Sm/RNP, IgG and its Fc and Fab fragments. Also shown are the IgM concentration and light chain expression of these clones. All supematants were tested undiluted.
Normal human monoclonal anti-Ro( SS-A) antibody
777
reactive clones varied from 0.1 to 44.9 pg/ml (mean 5.09 ug/ml). Either K (1 l/ 19) or h light (8/19) chains were expressed on the IgM cloned antibodies (Figure 1). These anti-Ro( SS-A) antibody-producing clones were stable for more than 2 years before being frozen for later use.
ELISA
inhibition
Inhibition of Ro(SS-A) binding was performed on all cloned IgM anti-Ro(SS-A) antibodies to assess their specificity. Since the results were similar for all clones, the inhibition results of only four clones are presented. These clones predominantly reacted to Ro( SS-A) since diluted supernatants lost their reactivity to La( SS-B) and Sm/RNP while maintaining relatively strong anti-Ro(SS-A) binding. Only antiRo( SS-A) reactivity remained in the dilutions selected for inhibition assays. Representative inhibition curves with cloned supernatants or with SLE serum IgM or IgG anti-Ro(SS-A) antibodies are shown in Figures 2 and 3, respectively. No inhibition of IgM monoclonal or serum polyclonal IgM or IgG anti-Ro(SS-A) binding was observed with RNA, IgG, ssDNA, DNA or BSA. Among the clones represented here (Figure 2), inhibition of anti-Ro(SS-A) binding was observed with Ro(SS-A) itself in a dose-dependent manner and ranged from 74 to 88Og inhibition with 25 ug of Ro(SS-A). Similar dose-dependent inhibition of anti-Ro(SS-A) binding to Ro(SS-A) was also achieved with 25 ug of La(SS-B) and/or Sm/RNP. Inhibition of polyclonal serum IgM anti-Ro( SS-A) binding of BRA andASH with Ro(SS-A) was, respectively, 70:/, and460/,, with La(SS-B) 57”” and 520/b, and with Sm/RNP 527,” and 28?, at 25 ug of antigen (Figure 3). The inhibition of the same two patients’ serum polyclonal IgG anti-Ro(SS-A) binding was achieved with lower with 5 pg), while no inhibition was concentrations of Ro(SS-A) (100 9/, inhibition observed with up to 25 ug of La( SS-B) or Sm/RNP.
Immunoblotting The Coomassie blue- and silver-stained SDS-PAGE as well as the Amido blackstained nitrocellulose showed that the Ro(SS-A) preparation contained a single band of 60 kDa. The immunoblotting (Figure 4) included as negative controls, HGM (lane l), a normal human IgM serum (lane 15) and an ELISA anti Ro( SS-A) negative supernatant BOU 11.4 (lane 2). All the anti Ro(SS-A) ELISA positive clones as well as an anti-Ro(SS-A) ELISA IgM positive serum TOW (lane 14), as a positive control, and two anti-Ro(SS-A) ELISA positive hybridomas (lanes 12 and 13) showed variable intensity of binding to the 60 kDa band. The same blot, developed with an anti-IgG alkaline phosphatase conjugate, revealed, as expected, binding to the 60 kDa band by the TOW serum only (data not shown). The La(SS-B) preparation displayed three bands: two major bands at 42 and 14 kDa and a minor band at 29 kDa. All but three clones (BOU 77.1, LAP 234.8 and 177.4), which showed anti-La(SS-B) ELISA reactivity, showed weak binding to one or all three bands of immunoblotted La(SS-B) (data not shown). Such binding of anti-La(SS-B) antibodies to La(SS-B) degradation products (29 and 14 kDa) has been noted by other investigators [34-361. Only the SLE sera showed binding to La(SS-B) when the blot was developed with an anti-IgG conjugate.
778
Helene Massicotte et al. LAP
156.4
LIL
loo’- l
Qr
16.13
100
70-m \ 90'
90
80 -
SO 70
60
50’
40
30. 20 IO t
”
IPg BOU
5Pa
25 PP
53.4
100 go-
90 -
60-
60-
70-
70 -
60 -
60-
50-
50-
40-
40‘
30-
30 -
20-
20 IO -
IO -
0’
’
1 #P
I
5 ILP
I
255~
Figure 2. Inhibition of cloned IgM anti-Ro(SS-A) antibody binding. One, 5 and 25 pg of the following inhibitors were added: Ro(SS-A) (0), La(SS-B) (A), Sm/RNP ( q), RNA (0), ssDNA ( l ), DNA (A), IgG ( n ) and BSA (4). The IgM concentration (pg of IgM/ml) and the dilution factor of each cloned supematant chosen to obtain an optical density of 1.0 were for BOU 53.4 (0.7 pg/ml) l/SO, BOU 11.4 (1.63 pg/ml) l/5, LAP 158.4 (2.96pg/ml) l/4 and LIL 16.13 (1.5 pg/ml) l/4.
None of the anti-Ro(SS-A) positive clones that bound to Sm/RNP by ELISA blotted any proteins associated with the Sm/RNP preparation (68kDa, A, BB’, C, D, E, F and G). Positive controls in this immunoblotting experiment consisted of clones selected specifically for anti-Sm/RNP binding and sera with anti-Sm/RNP reactivity (data not shown). The staining procedures used showed no cross-contaminating protein bands of Ro(SS-A), La(SS-B) or Sm/RNP by each other, or by IgG.
Normal human mon&lo&l BRA-IqM
._oI z
1u9
z m
8
anti-Ro(SS-A)
antibody
779
BRA-IqG t
IO*
5u9
25P9
0.1 ug
ASH-IqM
5P9
15co
ASH-IgG iooy-
IOOgoBO -
60504030 20 IO t
01
’
I
I
1 UP
5Pq
25~~
Figure 3. Inhibition of IgM anti-Ro(SS-A) binding of sera (dilution) BRA (l/80) and ASH (1 /SO) and of IgG anti-Ro(SS-A) binding of the same sera (l/10,000). The following inhibitors were added: Ro(SS-A) (Z), La(SS-B) (a), Sm/RNP (U), RNA (O), ssDNA (O), DNA (A), IgG (m) and BSA (*). The dilution of each serum was chosen to obtain an optical density of 1 .O.
Only those anti-Ro(SS-A) clones that bound IgG, Fab’ and/or Fc by ELISA showed binding to IgG components in immunoblotting experiments. These antibodies bound to both the heavy and light chain of IgG as well as to Fc and Fab’ fragments (data not shown). In immunoblots of normal lymphocyte lysates, only one monoclonal antibody, BOU 43.10, showed weak binding to bands of 79, 61, 42 and 13 kDa, which could correspond to Ro( S S-A) and La( S S-B). This could not be confirmed by inhibition studies because of the weak reactivity (data not shown). The same bands were also
780 Helene Massicotte et al.
Figure 4. Immunoblotting Ro(SS-A) with supernatants and sera. The amount of purified Ro(SS-A) (J. B. Harley) used per lane was 5ug. The undiluted cloned supematants (IgM concentration) or serum (dilution) used in each lane were as follows: 1: HGM, 2: BOU 11.4 (0.12 ug/ml), 3: BOU 53.6 (21.08 ug/ml), 4: BOU 77.1 (6.25 ug/ml), 5: LAP 234.5 (5.95 ug/ml) 6: BOU 318.1 (1.33 ug/ml), 7: LAP 177.4 (1.5 ug/ml), 8: LAP 193.1 (2.06 ug/ml) 9: LAP 149.2 (52.66 ug/ml) 10: LIL 13.12 (0.35 ug/ml), 11: BOU43.10 (6.35 ug/ml), 12: LIL 77 (3.28 ug/ml), 13: BOU 301 (5.84 ug/ml) 14: TOW (l/SO) 15: Normal serum (l/50). The F(ab’), goat anti-human IgM alkaline phosphatase conjugate (Jackson) and BCIP/NBT substrate (Gibco) were used to detect bound IgM.
seen when SLE sera with anti-Ro(SS-A) and anti-La(SS-B) reactivity were used. None of the other monoclonal antibodies showed binding to any bands in these cell lysates.
Immunoprecipitation Ten monoclonal IgM anti-Ro(SS-A) antibodies were tested for reactivity to the hYRNAs by immunoprecipitation with HeLa cell lysates. None of them precipitated hYRNAs. LIL 72.52 precipitated tRNA, 4.5s and 5s RNAs and a unique unidentified high molecular weight RNA; LIL 16.13 precipitated Ul and U2 RNAs and 7s RNA, consistent with its Sm/RNP ELISA reactivity (data not shown).
Discussion The present study shows that antibodies reactive with Ro(SS-A) can be successfully obtained by the technique of somatic cell hybridization. Anti-Ro(SS-A) antibodies
Normal human monoclonal anti-Ro(SS-A)
antibody
781
were seen in each of the three separate fusions performed (thus reproducible) employing lymphocytes from tonsils of three different normal individuals. This observation implies the existence of immunoglobulin genes encoding anti-Ro( SS-A) autoantibodies in normal humans. The present study further extends earlier observations from this laboratory on the generation of normal human:human hybridomas with autoreactivity [26]. This is also in keeping with the observations of others, that normals have the potential to produce natural antibodies to a variety of self antigens [37-40]. The binding properties of the normally-derived monoclonal IgM antiRo(SS-A) antibodies appear similar to the polyclonal serum IgM anti-Ro(SS-A) antibodies of patients with SLE. Monoclonal anti-Ro(SS-A) antibodies were selected by ELI SA for their reactivity to Ro(SS-A). Some of these monoclonal antibodies also reacted strongly with La(SS-B) or Sm/RNP, but not with RNA, BSA or caseine. However, the binding reactivity to La( SS-B) and Sm/RNP was weaker than binding to Ro(SS-A), since the former reactivity could be readily diluted out. Ro(SS-A), La(SS-B) and Sm/RNP were nonetheless all able to inhibit the Ro(SS-A) binding of these monoclonal anti-Ro(SS-A) antibodies in a dose-dependent manner at dilutions of antibody supernatant in which direct ELISA binding of La(SS-B) and Sm/RNP were undetectable. All of the monoclonal anti-Ro( SS-A) antibodies recognized a structural epitope on the 60 kDa Ro(SS-A) protein, and many also recognized a structural epitope on the polypeptides of the La(SS-B) protein. A presumed conformational epitope on Sm/RNI? proteins was recognized by some of the monoclonal antiRo(SS-A) antibodies, since those that reacted with Sm/RNP in ELISA failed to immunoblot Sm/RNP proteins. This may be explained by the observation that the intensity of binding reactivity to any antigen by ELISA or Western blotting does not always correlate [ 10,131, because these assays present the antigen in different forms to the antibodies. Thus, ELISA presents antigen as a native molecule in either a bound or free aqueous (inhibition) phase while Western blotting presents antigen in a linear form (denatured and reduced). None of these monoclonal anti-Ro(SS-A) antibodies immunoprecipitated hYRNAs, although one clone, which also bound Sm/RNP by ELISA, immunoprecipitated Ul and U2 RNA from HeLa cells. These results could be a reflection of the lack of sensitivity of the immunoprecipitation assay with HeLa cells by the relatively low concentration and affinity of the monoclonal IgM anti-Ro(SS-A) antibodies or because of lack of access to the epitope on the whole RNA-protein complex recognized by these particular monoclonal anti-Ro(SS-A) antibodies. Similar dose-dependent inhibition of Ro(SS-A) binding by Ro(SS-A), La(SS-B) and Sm/RNP was observed with the two representative polyclonal SLE serum IgM anti-Ro(SS-A) antibodies. In contrast, SLE serum IgG anti-Ro(SS-A) antibody binding was specifically inhibited with Ro(SS-A) and not with La(SS-B) or SmlRNP. These findings indicate that the property of polyreactivity is shared by patient serum IgM antibody and the normally-derived monoclonal IgM antibodies. These monoclonal anti-Ro( SS-A) antibodies also reacted to whole IgG or to its Fc and Fab fragments. This reactivity was clearly directed towards structural epitopes of the IgG molecule as revealed by the ability of these monoclonal IgM antiRo( SS-A) antibodies to blot both the heavy and light chains of IgG as well as its Fc and Fab’ fragments. However, since IgG, Fc or Fab’ could inhibit the Ro(SS-A)
782
Helene
Massicotte
et al.
binding, different sub-regions of the H and/or L chain may be involved in the reactivity of these antibodies to Ro( SS-A) and IgG. The binding to Ro( SS-A) could involve, for example, the VH chain region, while the rheumatoid factor activity, as recently described by Weisbart et al. [41], could involve the VL region chain. Particular H and L chain pairings can also alter the binding of the recombinant antibodies compared with the binding properties anticipated from the individual H and L chains [42]. Goldman et al. [43] described monoclonal rheumatoid factors that react with both the Fab and Fc region of autologous IgG. Anti-Fab antibodies have also been frequently noted amongst patients with rheumatoid arthritis [44] and in some instances closely associated with the presence of antibody reactivity to Ro(SS-A) [45]. Indeed, Ro(SS-A) and IgG Fab appear to share epitopes that are bound by polyclonal serum IgG anti-Ro(SS-A) antibody [45]. Further, it has been shown in mice that monoclonal antibodies selected for rheumatoid factor activity have polyspecificity to other unrelated autoantigens [46]. In a previous study from this laboratory, normal human monoclonal IgM antibodies selected for DNA binding frequently revealed polyreactivity to a variety of nucleic acid antigens, including DNA, cardiolipin and synthetic polynucleotides with different base structures [26], but none showed binding to the RNA-protein associated antigens (Ro(SS-A), La(SS-B) and Sm/RNP; data not shown). Similarly, the antibodies selected here for their reactivity to Ro(SS-A) also showed polyreactive binding to other unrelated RNA-protein associated antigens, but rarely to DNA related antigens (2/19 anti-Ro(SS-A) antibodies bound DNA) and uncommonly to cardiolipin (7/19). This dichotomy parallels the observations in patients with where anti-RNA-protein-associated IgG polyclonal autoantibody response, autoantibodies and anti-DNA autoantibodies often appear independently. The present study does not explain why these monoclonal IgM antibodies bound to what appear to be structurally unrelated RNA-protein complexes. There is no known shared structural epitope on Ro( SS-A) and La( SS-B) recognized at present, however, Ro(SS-A) and La(SS-B) are often found as a complex [20, 21, 23, 241. Moreover, a significant proportion of patients have serum IgG antibodies that bind to Ro(SS-A) as well as to La(SS-B) [2, 11, 12,20-241. It is conceivable, therefore, that the normal immune system may have B cells whose Ig receptor recognizes Ro(SS-A) and La(SS-B) individually, as a complex, or in some instances may recognize some epitope common to both of these molecules. In summary, the present study reveals that normal human B cells have the genetic potential to produce antibodies to Ro(SS-A). Some monocional anti-Ro(SS-A) antibodies also strongly bound by ELISA to La( SS-B), Sm/RNP, IgG, Fab’ and Fc. Immunoblotting and inhibition studies disclosed reactivity of the same IgM molecules to both Ro(SS-A) and La(SS-B) proteins. This suggests the possibility that there is either a common shared protein epitope on the Ro(SS-A) and La(SS-B) antigens or that different subsites exist for the Ro(SS-A) and La(SS-B) antigens within the same antibody variable region. In contrast, the reactivity of these antibodies to Sm/RNP by ELISA may reflect binding to conformational determinants, since these antibodies did not reveal binding to Sm/RNP proteins by immunoblotting. It remains to be determined whether the normally-derived IgM monoclonal anti-Ro(SS-A) autoantibodies and the polyclonal serum IgM anti-
Normal human monoclonal anti-Ro( SS-A) antibody 783 Ro(SS-A) autoantibodies from SLE patients, which have similar properties, are derived from and share similar germline immunoglobulin V region genes. It is also controversial whether or not natural antibodies, which appear in normal individuals, are derived from the same precursor B cells that give origin to the IgG antibodies which either are expressed in disease or have pathogenic potential [47, 481. An analysis and comparison of the variable regions of the normal IgM anti-Ro(SS-A) antibodies produced by these hybridoma B cells and the IgG anti-Ro(SS-A) autoantibodies produced in disease should provide some insight into the factors responsible for determining their relatedness, expression and regulation.
Acknowledgements Dr Massicotte’s work was supported by a postdoctoral Fellowship from the Canadian Arthritis Society. This work is supported by a research grant from the Canadian Arthritis Society. Dr Harley is supported by the NIH and the Veterans Administration.
References 1. Lerner, M. R. and J. A. Steitz. 1979. Antibodies to small nuclear RNAs complexed with proteins are produced by patients with systemic lupus erythematosus. Proc. Natl. Acad. Sci. USA 76: 5495-5499 2. Harley, J. B., E. L. Alexander, W. B. Bias, 0. F. Fox, T. T. Provost, M. Reichlin, H. Yamagata, and F. C. Arnett. 1986. Anti-Ro/SS-A and anti-La/SS-B in patients with Sjogren’s Syndrome. Arthritis Rheum. 29: 196-206 3. Watson, R. M., A. T. Lane, N.K. Bamett, W. B. Bias, F. C. Amett, and T. T. Provost. 1984. Neonatal lupus erythematosus: a clinical, serological and immunogenetic study with review of the literature. Medicine (Baltimore) 63: 362-378 4. Lee, L. A., C. E. Harmon, J. C. Huff, D. A. Norris, and W. L. Weston. 1985. The demonstration of SS-A/Ro antigen in human fetal tissues and in neonatal and adult skin. J. Invest. Dermatol. 85: 143-146 5. Deng, J. S., R. D. Sontheimer, and J. N. Gilliam. 1985. Expression ofRo/SS-A antigen in human skin and heart.J. Invest. Dermatol. 85: 412-416 6. Harmon, C. H., J. S. Deng, C. L. Peebles, and M. E. Tan. 1984. The importance of tissue substrate in the SS-A/Ro antigen-antibody system. Arthritis Rheum. 27: 166-173 7. Wolin, S. L. and J. A. Steitz. 1983. Genes for two small cytoplasmic Ro RNAs are adjacent and appear to single copy in the human genome. Cell 32: 735-744 8. Kato, N., H. Hoshino, and F. Harada. 1982. Nucleotide sequence of 4.5s RNA (C8 or HY5) from HeLa cells. Biochem. Biophys. Res. Commun. 108: 363-370 9. Ben-Chetrit, E., E. K. L. Chan, K. F. Sullivan, and E. M. Tan. 1988. A 52 kd protein is a novel component of the SS-A/Ro antigenic particle. J. Exp. Med. 167: 1560-1571 10. Rader, M. D., C. O’Brien, Y. Liu, J. B. Harley, and M. Reichlin. 1989. Heterogeneity of the Ro/SS-A antigen: different molecular forms in lymphocytes and red blood cells. J. Gin. Invest. 83: 1293-1298 11. Woiin, S. L. and J. A. Steitz. 1984. The Ro small cytoplasmic ribonucleoprotein: identification of the antigenic protein and its binding site on the Ro RNAs. Proc. Nacl. Acad. Sci. USA 81: 1996-2000 12. Lieu, T. S., M. Jiang, J. C. Steigerwald, and E. M. Tan. 1984. Identification of the SS-A/Ro intracellular antigen with autoimmune sera.g. Zmmunol. Methods 71: 217-228 13. Yamagata, H., J. B. Harley, and M. Reichlin. 1984. Molecular properties of the Ro/SS-A antigen and enzyme-linked immunosorbent assay for quantitation of antibody. J. Clin. Invest. 74: 625-633
784
Helene
Massicotte
et al.
14. McNeilage, L. J., S. Whittingham, and I. R. MacKay. 1984. Autoantibodies reactive with small ribonucleoprotein antigens: a convergence of molecular biology and clinical immunology. 3. Clin. Lab. Zmmunol. 15: 1-17 15. Reddy, R., D. Henning, E. M. Tan, and H. Busch. 1983. Identification of a La protein binding site in a RNA polymerase III transcript (4.5 IRNA). 3. Biol. Chem. 258: 8352-8356 16. Madore, S. J., E. D. Wieben, and T. Pederson. 1984. Eukaryotic small ribonucleoproteins.3. Biol. Chem. 259: 1929-1933 17. Kurilla, M. G. and J. D. Keene. 1983. The leader RNA of vesicular stomatitis virus is bound by a cellular protein reactive with anti-La lupus antibodies. Cell 34: 837-845 18. Habets, W. J., J. H. den Brok, A. M. Th. Boerbooms, L. B. van de Putte, and W. J. van Venrooij. 1983. Characterization of the SS-B (La) antigen in adenovirus-infected and uninfected HeLa cells. EMBOJ. 2: 1625-1631 19. Francoeur, A. M., E. K. L. Chan, J. I. Garrels, and M. B. Mathews. 1985. Characterization and purification of lupus antigen La, an RNA binding protein. Mol. Cell. Biol. 5: 586-590 20. Eisenberg, R. A. 1985. Association between Ro and La antigenic determinants: immunodiffusion analysis of human spleen extract. 3. Zmmunol. 3 5: 1707- 17 13 21. Hardin, J. A. 1986. The lupus autoantigens and the pathogenesis of systemic lupus erythematosus. Arthritis Rheum. 29: 457-460 22. Boire, G. and J. Craft. 1989. Biochemical and immunological heterogeneity of the Ro ribonucleoprotein particles. Analysis with sera specific for Rohy5 particle. 3. Clin. Invest. 84: 270-279 23. Boire, G. and J. Craft. 1990. Human Ro ribonucleoprotein particles: characterization of native structure and stable association with the La polypeptide. 3. Clin. Invest. 85: 1182-1190 24. Mamula, M. J., E. D. Silverman, R. M. Laxer, L. Bentur, B. Isacovics, and J. A. Hardin. 1989. Human monoclonal anti-La antibodies. The La protein resides on a subset of Ro particles. 3. Zmmunol. 143: 2923-2928 25. Cairns, E., J. St Germain, and D. A. Bell. 1985. The in vitro production of anti-DNA antibody by cultured peripheral blood or tonsillar lymphoid cells from normal donors and SLE patients. 3. Immunol. 13: 3839-3844 autoantibody-producing 26. Cairns, E., J. Block, and D. A. Bell. 1984. Anti-DNA hybridomas of normal human lymphoid cell origin. 3. C/in. Invest. 74: 880-887 27. Gaither, K. K., 0. F. Fox, H. Yamagata, M. J. Mamula, M. Reichlin, and J. B. Harley. 1987. Implications of anti-Ro/Sjogren’s Syndrome A antigen autoantibody in normal sera for autoimmunity. 3. Clin. Invest. 79: 841-846 28. Harley, J. B., H. Yamagata, and M. Reichlin. 1984. Anti-La/SS-B antibody is present in some normal sera and is coincident with anti-Ro/SSA precipitins in systemic lupus erythematosus.3. Rheum. 11: 309-314 29. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the bacteriophage T4. Nature 227: 680-685 30. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoresis transfer of protein from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76: 4350-4360 31. Forman, M. S., M. Nakamura, T. Mimori, C. Gelpi, and J. A. Hardin. 1985. Detection of antibodies to small nuclear ribonucleoproteins and small cytoplasmic ribonucleoproteins using unlabeled cell extracts. Arthritis Rheum. 28: 1356-I 361 32. Massicotte, H., J. Rauch, Y. Shoenfeld, and H. Tannenbaum. 1984. Influence of fusion cell ratio and cell plating density on production of human-human hybridomas secreting anti-DNA autoantibodies from patients with systemic lupus erythematosus. Hybridoma 3: 215-222 33. Coller, H. A. and B. S. Coller. 1983. Statistical analysis of repetitive subcloning by the limiting dilution technique with a view toward ensuring hybridoma monoclonality. Hybridoma 2: 91-96
Normal human monoclonal anti-Ro( SS-A) antibody
785
of 34. Venables, P. J. W., P. R. Smith, and N. Maini. 1985. Purification and characterization the Sjogren’s Syndrome A and B antigens. Clin. Exp. Immunol. 5: 731-738 35. Harley, J. B., M. 0. Rosario, H. Yamagata, 0. F. Fox, and E. Koren. 1985. Immunologic and structural studies of the lupus/Sjogren’s Syndrome autoantigen, La/SSB, with a monoclonal antibody. J. Clin. Znvesr. 76: 801-806 and E. M. Tan. 1986. Epitopes, structural domains, 36. Chan, E. K. L., A. M. Francoeur, and assymmetry of amino acid residues in SS-B/La nuclear protein. 3. Zmmunol. 136: 3744-3749 B. G., G. Dighiero, and S. Avrameas. 1983. Naturally occuring antibodies 37. Guilbert, against nine common antigens in human sera. 3. Immunol. 128: 2779-2787 R. G. Whalen, 38. Dighiero, G., P. Lymberi, J.-C. Mazie, S. Rouyre, G. S. Butler-Browne, and S. Avrameas. 1983. Murine hybridomas secreting natural monoclonal antibodies reacting with self antigens. 3. Immunol. 131: 2267-2272 the “know thyself’ of 39. Avrameas, S. 1986. Natural autoreactive B cells and autoantibodies: the immune system. Ann. Zmmunol. Inst. Pasteur 137D: 150-156 D. M., R. A. Eisenberg, and A. D. Steinberg. 1990. Development of the 40. Klinman, autoimmune B cell repertoire in MRL-lpr/lpr mice. 3. Zmmunol. 144: 506-571 41. Weisbart, R. H., A. L. Wong, D. Noritake, A. Kacena, G. Chan, C. Ruland, E. Chin, I. S. Y. Chen, and J. D. Rosenblatt. 1991. The rheumatoid factor reactivity of a human IgG monoclonal autoantibody is encoded by a variant VkII L chain gene.3. Zmmunol. 147: 2795-2801 42. Radic, M. Z., M. A. Mascelli, J. Erikson, H. Shan, and M. Weigert. 1991. Ig H and L chain contributions to autoimmune specificities.3. Zmmunol. 146: 176-182 M., J. C. Renversez, and I’. H. Lambert. 1983. Pathological expression of 43. Goldman, idiotypic interactions: immune complexes and cryoglobulins. Springer Semin. Immunopathol. 6: 33-49 of 44. Nasu, H., D. Chia, Y. Iwaki, E. V. Barnett, and P. I. Terasaki. 1981. Cytotoxicity anti-Fab antibodies against B lymphocytes in rheumatoid arthritis. Arthritis Rheum. 24: 1278-1282 45. Mamula, M. J., 0. F. Fox, H. Yamagata, and J. B. Harley. 1986. The Ro/SS-A autoantigen as an immunogen: some anti-Ro/SS-A antibody binds IgG. 3. Exp. Med. 86: 1889-1901 46. Rubin, R. L., R. S. Balderas, E. M. Tan, F. J. Dixon, and A. N. Theofilopous. 1984. Multiple autoantigen binding capabilities of mouse monoclonal antibodies selected for rheumatoid factor activity. 3. Exp. Med. 159: 1429-1440 1987. The genetic origin of 47. Kofler, R., F. J. Dixon, and A. N. Theofilopoulos. autoantibodies. Immunol. Today 8: 374-48 48. Bona, C. A. 1988. V genes encoding autoantibodies: molecular and phenotypic characteristics. Annu. Rev. Immunol. 6: 327-358
