Article No. jaut.1999.0295, available online at http://www.idealibrary.

com on Journal of Autoimmunity (1999) 13, 95–102

Screening Tests for Antinuclear Antibodies (ANA):

Selective Use of Central Nuclear Antigens as a Rational
Basis for Screening by ELISA

Eli Olaussen1 and Ole Petter Rekvig1,2

1
Department of Immunology, University Rational, fully automated ELISA screening tests for ANA have become
Hospital of Tromsø, N-9038, Tromsø, accessible for routine use. Full length, recombinant or purified naturally
Norway expressed antigens can be selected to deliver well defined screening tests,
2
Department of Molecular Genetics, designed to detect ANAs of established clinical significance, while unknown
Institute of Medical Biology, University of specificities with no clear biological impact are omitted. By comparing results
Tromsø, N-9037 Tromsø, Norway of ANA screening by immunofluorescence (IIF) on tissue section with HEp-2
cells, a correlation of 88% was revealed, while 86% of HEp-2-positive sera
bound tissue sections. Eighty-nine percent of HEp-2 ANA-positive sera bound
in ANA ELISA I. Seventy-five percent of ANAs detected in ELISA I also bound
in ELISA II. These correlations were statistically significant. ANAs detected in
ANA ELISA I were all detected upon retesting in both ANA screening ELISA
systems, and also by antigen-specific ELISAs. Specific ANAs defined by
ELISA, immunodiffusion or Crithidia tests were detected in both ANA ELISA
tests. These data demonstrate that most ANAs detected by immunofluores-
cence tests contain a dominating repertoire of specificities covered by the ANA
ELISA systems. The sensitivity of the different ELISA tests can easily be
Key words: antinuclear adjusted to internationally defined consensus levels, and screening for ANA
antibodies, autoimmunity, using both ELISA systems detects expected ANA specificities in ongoing
ELISA, immunofluorescence international quality assessment programs. © 1999 Academic Press

Introduction molecules, as indicated for antibodies to dsDNA by

Antinuclear antibodies (ANA), as a distinct group of
autoantibodies, are defined as those recognizing anti-
gens morphologically confined to the cell nucleus.
Because they react with such antigens irrespective of
species origin, and also human origin, they are con-
ceptually defined as true autoantibodies. However, it
is not established that the autologous antigens always
account for their production, or whether such anti-
bodies are induced as a consequence of stimulation
with heterologous antigens mimicking autologous
antigens (molecular mimicry, see for example refer-
ences [1, 2]), or complexes between autologous and
heterologous antigens [3–5] or by other events that
may contribute to, but not as a sole factor account for,
increased antibody production, such as polyclonal B
cell stimulation [6]. However, true autoantibodies
may also be induced by autoimmune B-cells and
T-cells, both recognizing determinants on single
autoantigens or complexes of different autologous
Correspondence to: Ole Petter Rekvig, Department of Molecular
Genetics, Institute of Medical Biology, University of Tromsø, N-9037
Tromsø, Norway. E-mail: olepr@fagmed.uit.no
95
0896–8411/99/050095+08 $30.00/0
the pioneering results of Datta et al. [7, 8]. The molecu-
lar mechanisms accounting for such antibodies are all,
however, compatible with affinity maturation, result-
ing in high affinity IgG antibodies; with the important
practical consequence that tests for most antinuclear
antibodies should be limited to detection of IgG
antibodies.
Numerous different ANAs are described, reactive
with nuclear antigens such as nucleosomal antigens
(DNA and the histone classes H1, H2A, H2B, H3 and
H4), enzymes and regulatory proteins linked to tran-
scription, DNA repair, and others [9]. Although more
than several 10-folds of specificities have been
described, only a few have proven to be clinically
important and are thus implemented in regular
diagnostic protocols [9, 10]
Certain ANAs may relate to a distinct syndrome,
defining its diagnostic value [9, 10], but this does not
necessarily imply that such specific ANAs are
involved in the pathophysiological process of a given
disease. Others again may have such an impact, as
may be the case for at least some populations reactive
with dsDNA (reviewed in reference [11]). However,
subgroups of anti-dsDNA antibodies may also be
detected in patients without signs of active disease,
© 1999 Academic Press
96
indicating that a positive ANA test and detection of
specific ANAs should be interpreted critically.
Another important problem that reduces the value of
a non-selective ANA detection derives from the
fact that both ANA (as a non-specified group of
autoantibodies), and also specific ANAs may be
produced after ingestion of a relatively wide variety
of drugs (reviewed in reference [12]), and not
infrequently among normal, particularly elderly, indi-
viduals [13, 14], or linked to certain infections [15, 16].
These latter facts may reduce the impact of ANA
detection as helpful diagnostic tools. We therefore feel
it is wise to perform the ANA test as a restricted
screening test, allowing the detection of preselected
specificities of established diagnostic and/or patho-
physiological significance. While the non-selective
detection of any ANA by indirect immunofluorescence
(IIF) may inform as to the ‘autoimmune status’ of a
given patient, it is not justified to use the test for that
purpose, due to the poor clinical value of unclassified
ANAs. Another approach would be to introduce tests
that detect a stringent repertoire of ANA specificities of
established clinical significance. Thus, the latter screen-
ing approach aims to diagnose syndromes known to
include production of such distinct ANA specificities.
In this report, we compare two IIF tests with each
other, and with two different ANA screening ELISAs.
The results demonstrate that a large proportion of the
ANA specificities detected by IIF are also detected by
the ANA ELISA screening systems and include anti-
bodies to SSA, SSB, RNP, Sm, Jo-1, Scl-70, centromer,
DNA and also potentially to histones (not tested in
this study). These results strongly indicate that among
all antinuclear antibodies detected on routine bases
with either of the test systems described above, few
bind antigens outside the selected repertoire of the
ANA ELISA plates used in this study.
Materials and Methods
Sera
Sera sent to our laboratory for screening of ANA were
included in the present study. For the purpose of
screening, random sera were tested in the order they
were received. In certain experiments, HEp-2 ANA-
positive and-negative or distinct antigen-specific sera
were selected for assessing the quality of the two ANA
screening ELISA systems (see below). The numbers
of sera included in each of the different tests are
indicated in each Table or Figure, and differ between
experiments.
Indirect immunofluorescence (IIF) for detection of
ANA
Tissue sections of mouse kidney/mouse stomach
were used for IIF (referred to below as IIF on tissue
sections), 4 ìm thick freeze-sections were prepared,
dried and subjected to indirect immunofluorescence
E. Olaussen and O. P. Rekvig
as described previously [17]. Patients’ sera were
screened at serum dilution 1/50 in PBS (0.2% BSA,
PBSA). Bound antibodies were detected by using
FITC-conjugated antibodies specific for human IgG
(Dako, Copenhagen, Denmark). Indirect immuno-
fluorescence using HEp-2 cells as substrate was
carried out as described [18].
Screening ELISA systems for detection of ANA
Two commercially available ANA screening ELISA
systems were included in these investigations and
compared with each other and with an IIF test using
HEp-2 cells as substrate.
The first system, Diastat, denoted in this report as
ANA ELISA I, was obtained from Shield Diagnostic
Ltd. (Dundee, Scotland). The test was performed
exactly as recommended in the kit information. The
cut-off value, determined as recommended, was
slightly modified according to our own results, and
was set to a ratio (OD test serum/OD positive refer-
ence serum) of 0.9. However, for defining ANA
specificities for Jo-1, Scl-70 and centromer antigens, a
cut-off of 0.6 was used, as these specificities gave
consistently lower values in the ANA test compared to
the other specificities (data not shown).
The other system, ANA Rheuma Screen, here
referred to as ANA ELISA II, was obtained from
Pharmacia & Upjohn, Diagnostics GmbH & Co. KG
(Freiburg, Germany). The test protocol used in this
study was as recommended by the manufacturer. A
serum was considered positive at a ratio ≥1.4, while
borderline reactions (1<ratio<1.4) were considered
negative when the purpose was to compare actual
results with those obtained in other test systems.
ELISA automation
The different ELISA tests protocols were applied to
the fully automated Behring ELISA Processor III
(Behringwerke, Marbourg, Germany), and ELISA
results reported here are generated and calculated
using this equipment.
Statistics
Degrees of correlation between the different test
results as factors were determined by regression
analyses using the Pearson correlation coefficient.
Results
Correlation between two IIF techniques using
tissue sections or HEp-2 cells as substrate for
detecting ANA
In order to investigate the degree of correlation
between two classical IIF tests (tissue sections vs.
Screening of ANA by ELISA 97

Table 1. Comparison of different ANA screening test systems. ANA detected in each of the screening tests are analysed for
reproducibility in the other screening tests

Number (%) of ANA-positive sera upon retesting by:

Number of
Serum ANA defined by: ANA ELISA
sera IIF HEp-2 cells IIF tissue sectionsa
Ib IIc

IIF tissue sectionsa 117 103 (88%)

IIF on HEp-2 cells 57 49 (86%)
IIF on HEp-2 cells 123 110 (89%)
ANA ELISA I 81 70 (86%)
ANA ELISA Ib 173 130 (75%)

a
IIF on tissue sections refers to indirect immunofluorescensce using sections of mouse kidney/stomach as substrate.
b
ANA ELISA I refers to ANA screening ELISA obtained from Shield.
c
ANA ELISA II refers to ANA screening ELISA obtained from Pharmacia & Upjohn.

Table 2. Comparison of ANA screening methods: highly compared to the high number of different ligands in
significant correlation between ANA screening by IIF using intact cell nuclei. The repertoire of ANAs detected by
HEp-2 cells as substrate vs. the use of ANA ELISA I ANA ELISA I, therefore, seems to represent a domi-
nating spectrum of antibodies in randomly selected
ANA ELISA I patients’ sera.
Positive Negative

Positive 110 13 Central ANA specificities are detected in ANA

IIF HEp-2
cells
Negative 27 471
a
P<0.0001.
HEp-2 cells as substrates), 117 sera reacting with
nuclei presented by tissue sections were selected. Of
these 117 sera, 103 (88%) reacted with nuclei of HEp-2
cells (Table 1). Conversely, of 57 HEp-2 ANA-positive
sera, 49 (86%) reacted upon retesting with nuclei of
tissue sections (Table 1). Thus, by determining the
correlation bi-directionally, more than 85% (but less
than 90%) of the ANAs reacted in both IIF test systems
upon retesting.
Correlation between IIF using HEp-2 cells with
ANA ELISA I for detection of ANA
In the next series of experiments, we tested whether
ANAs detected in the HEp-2 assay could be detected
upon retesting in the ANA ELISA I. Of 123 HEp-2
ANA-positive sera, 110 (89%) could be detected in this
ANA screening ELISA (Table 1), while of 81 ANA
ELISA I-positive sera, 70 (86%) bound in the HEp-2
assay (Table 1). Similarly, as demonstrated in Table 2,
of 498 HEp-2 ANA-negative sera, only 27 tested
positive in the ANA ELISA I, while of 484 ANA ELISA
I-negative sera, only 13 were positive upon retesting
in the HEp-2 assay. Thus, the ANA results obtained in
the two principally different assays using the same
sera correlated in a highly significant way (P<0.0001).
This is a surprisingly strong correlation, as the
number of antigens in the ANA ELISA coat is limited,
screening by ELISA
To verify that the ELISA screening for ANA enables
detection of central ANA specificities, 105 sera con-
taining antibodies to SSA, SSB, Sm or RNP as deter-
mined by double immunodiffusion, and 41 Crithidia
luciliae anti-dsDNA antibody-positive sera were
re-analysed by the ANA ELISA I. All sera contained
antibodies interacting with ligands present in the
coated wells (data not shown). Since these antibodies
are selected by relatively insensitive techniques com-
pared to ELISA (immunodiffusion or Crithidia tests)
the latter results confirm that at least strong specific
ANAs are detected in this ANA ELISA.
Correlation between two different ANA screening
ELISA systems for detection of ANA
In the next series of investigations, selected ANA-
positive and-negative sera (a total of 423) were tested
in parallel using ANA ELISA I (Shield) and ANA
ELISA II (Pharmacia & Upjohn). These two ELISA
systems have an almost identical composition of
ligands in their coated wells. However, the former
contain mostly immunopurified naturally produced
ligands, while in the latter mostly recombinantly pro-
duced ligands are used. As demonstrated in Figure 1,
a strong correlation was observed between results
obtained using the same sera in the two tests. How-
ever, more than 20% of the ANA-positive sera reacted
in only one of the tests (Table 1): 173 sera tested
positive in the ANA ELISA I, while only 130 of these
173 sera were positive in the ANA ELISA II. Among
the 423 selected ANA ELISA I-positive or -negative
sera tested in both ANA screening ELISAs, 130 were
98 E. Olaussen and O. P. Rekvig

10

8
ANA ELISA II
7 Ratio ≥ 1.4
+ –

ANA ELISA II
6

ANA ELISA I
ratio ≥ 0.9
+ 130 43
5
– 8 242
4

2 p < 0.001

0 1 2 3 4 5 6 7 8 9 10
ANA ELISA I
Figure 1. Comparison of results obtained by ANA ELISA I (x-axis) with those obtained by ANA ELISA II (y-axis). For this
comparison, 173 ANA ELISA I-positive and 250 ANA ELISA I-negative sera were selected and retested in ANA ELISA II. The
inserted 2×2 table for comparison of the results presented in the plot demonstrates a highly significant correlation between
the tests (P<0.001)

Table 3. ANA specificity determination of ANA detected by detected. The distribution and numbers of specificities
ANA screening ELISA I and II in these latter 35 sera are presented in Figure 2. In
Figure 3, results are presented for 780 sera tested in
Number of Number of the ANA ELISA II. The results of the 59 ANA-positive
Method ANA-positive ANA with % ‘Hits’ sera with regard to fine specificities vs. strength of the
sera defined specificityb
ANA ELISA signals indicate that there is no apparent
correlation between any distinct ANA specificity and
ANA ELISA Ia 626 378 60 strength of the ANA test result of a given serum
ANA ELISA IIa 59 35 59 (Figure 3). This was not established with serum anti-
bodies to Jo-1 during this part of the investigation.
a
Composition of the ligands used in these ANA screening ELISAs Sera with multiple specificities tended to give stronger
are viturally identical, and compromise DNA, histones, SSA, SSB, ANA ELISA signals than monospecific sera (Figure 3).
RNP, Sm, Jo-1, Scl-70 and centromer.
b
Specificity of ANA-positive sera is established by using ELISA
plates individually coated with each of the antigens present in the
ANA screening wells, except for histones which were not How to determine amounts of a given antibody
included in these investigations.
In order to quantify antinuclear antibodies, three
different approaches are widely used. Except for
double-positive while 242 were double-negative for serum titration, which requires the use of several
ANA (P<0.0001; see Figure 1, inserted 2×2 table). wells, results (as an optical density) obtained from a
single serum dilution is often related to a positive
standard reference serum (as a simple ratio), or to a
standard curve (as a unit). Whether values given as
Specificity of ANAs detected in ANA screening ratio could substitute for values (units) obtained by
ELISA II standard curves was tested. By comparing these two
factors observed for antibodies to dsDNA (Figure 4A)
ANA-positive sera detected by ANA ELISA II were or for thyroid peroxidase (TPO) (Figure 4B), a very
examined further to define their specificities. Specific strong correlation between the two values was indeed
ELISAs were performed, with a repertoire corre- observed (r=0.938 and r=0.913 for anti-TPO and anti-
sponding to the antigen repertoire in the ANA screen- dsDNA antibodies respectively).
ing ELISAs, except that the histone specific ELISA was
not included in this study. As shown in Table 3, of 626
ANA-positive sera as determined by ANA ELISA I, Discussion
378 (60%) sera could be shown to contain distinct
ANA specificities. For ANA ELISA II, of 59 ANA- Our knowledge relating to the molecular and cellu-
positive sera, in 35 (59%) specific ANAs could be lar origins of antibodies detected in autoimmune
Screening of ANA by ELISA 99

40

35

30

25
Number

20

15

10

0
ANA-positive SSA SSB Sm RNP Jo 1 Scl Centr DNA
sera Number of ANA specificities classified by specific
Pharmacia & Upjohn ELISAs
Figure 2. Distribution and number of antigen-specific ANAs detected by screening of 780 sera for ANA using ANA ELISA II.
Of 59 ANA-positive sera, 35 reacted in the indicated antigen-specific ELISA tests (also see Figure 3 for details).

7 SSA
SSB SSA
Sm SSB
6 RNP Sm RNP
DNA
CL
Cent
Cent
ANA ELISA II Ratio

5
SSA SSA SSA
RNP SSA SSB SSB
4 SSB SSB DNA
SSB Cent SSA
DNA Scl DNA
DNA
3 CL SSA
DNA SSA DNA
SSA DNA
SSA CL DNA CL
Scl SSB
2 Cent DNA Scl Sm DNA
Scl Sm DNA DNA Scl
DNA DNA
1

0 100 200 300 400 500 600 700 800

Number of tested serum samples (n = 780)
Figure 3. Of 780 sera tested, 59 were ANA-positive when screened in ANA ELISA II. The individual specificities of these
sera for selected nuclear antigens are indicated, and the strength of the ANA screening reaction (as ratio) is plotted against
these specificities. As indicated, there is a tendency for stronger binding in the ANA ELISA II for sera with multiple
ANA specificities.

syndromes are steadily evolving (for example see
references [7, 8, 19–21] for antibodies to DNA). In
general, since ANA screening is mostly aimed at, and
designed for, the detection of IgG antibodies, such
antibodies are presumably (auto-) antigen driven, T
helper cell-dependent immune responses. Further-
more, diagnostic tests for determining autoimmune
syndromes are continuously improving both at the
analytical and the technical level. This new analytical
insight may have important conceptual consequences.
For example, in the context of testing for antinuclear
antibodies, new test principles for screening of this
group of autoantibodies have been developed aimed
at diagnosing ANA specificities that the scientific
community has determined as clinically central (for
example see references [9, 10] and the ANA ELISA
kit inserts obtained from Shield Diagnostics and from
Pharmacia & Upjohn). Such new tests are now com-
mercially available. Until recently, the presence of
serum antinuclear autoantibodies has been analysed
using cell nuclei of different origins [17, 18], and
positive tests are further defined by specific immu-
noassays such as immunodiffusion and more recently
by specific ELISA systems. However, the only analyti-
cal limitation of ANA specificities detected by this
approach is the number of available antigens in the
nucleus. Thus, the possible presence of unclassified
specificities we have no specific tests for, and we
100
A B
250
200
150
Units
100
50
r = 0.913
0 1 2 3 4 5 6 7
Ratio
Figure 4. Correlation between ratio and units as quantitative values for antibodies to dsDNA (A) or to TPO (B)
(autoantibodies to thyroperoxidase).
therefore have no clinical insight into, will also inter-
fere with such screening tests. This is an analytical
drawback, as, in a rational sense, ANA as a phenom-
enon reflects the presence of undefined arrays of
antibodies, some of which have no proven value for
autoimmune diseases. A consequence of these consid-
erations would be to converge the spectrum of ANA
specificities towards a preselected repertoire of ANAs
with established clinical impact. Combining knowl-
edge gained over the last decades, there seems to exist
a consensus core repertoire of specific ANAs consist-
ing of antibodies against nucleosomal antigens (DNA
and histones, [22]) and extractable nuclear antigens
(ENA) (e.g. SSA, SSB, RNP, Sm, Jo-1 Scl-70 and
centromer, [9, 10]). In fact, both Shield Diagnostics and
Pharmacia & Upjohn have recently developed ANA
screening ELISA systems largely aimed at detection of
a limited and selected spectrum of antibodies. Here,
we present results deriving from our continuous
evaluation program for analytical tests, using ANA
screening ELISA systems from Shield Diagnostics and
from Pharmacia & Upjohn, both implementing the
repertoire of nuclear antigens given above. The main
differences between these ELISA kits refers to the
source of the antigens and to the immunoglobulin
classes detected; both factors that may influence the
results when comparing the two tests. While ANA
ELISA I (Shield) uses immunopurified antigens from
cells, the system developed by Pharmacia & Upjohn
(ANA ELISA II) uses mostly recombinantly produced
antigens (see kit information). Furthermore, in the
ANA ELISA I, both IgM and IgG antibodies are
detected, while only IgG antibodies are detected in the
ANA ELISA II.
By bi-directional testing and comparison, we found
that 85–90% of the serum ANAs bound in both
immunofluorescence assays (IIF on tissue sections vs.
HEp-2, and vice versa). Thus, 10–15% of the antibodies
were selectively detected by one or the other IIF test,
indicating a difference in specificity rather than in
sensitivity of the tests. Indeed, many functional
nuclear components are cell-specific, such as certain
E. Olaussen and O. P. Rekvig
1000
800
600
Units
400
200
r = 0.938
0 1 2 3 4 5 6 7
Ratio
transcription factors [23], whereas others are shared
by virtually all dividing cells, such as components of
nucleosomes.
By testing HEp-2-positive sera in ANA ELISA I, up
to 90% of the sera bound in the latter assay, although
the repertoire of antigens present in the ELISA is
highly restricted compared to the theoretically high
numbers of reactive ligands in the HEp-2 cell nucleus.
The ANA ELISA I was found useful as a screening test
for the following reasons. Selected sera containing
diverse specific ANAs directed against ligands
present in the ANA ELISA wells, all bound. Further-
more, up to 60% of all ANAs detected in the screening
ELISA possessed specificities that could be deter-
mined by subsequent specific ELISA tests. When
applying specific tests on ANA detected by IIF, the
frequency of ‘hits’ may even be lower (data not
shown). The reason why 40% of the ANA-positive
sera contained specificities that we were unable to
determine may be due to the fact that a specific test for
histone antibodies, which may be detected in the
screening test, was not included in this study.
Secondly, the screening by ANA ELISA I is designed
to detect IgG and IgM antibodies, while the specific
tests, except for dsDNA, only detect IgG antibodies.
Alternatively, there may be some differences in the
sensitivity of specific ELISA tests compared to the
ANA screening ELISA.
As ANA detection by the HEp-2 test correlated
strongly with that for ANA ELISA I, the two ANA
ELISA screening systems correlated to a lower degree,
although still highly significantly (see Table 1). Thus,
more ANAs were selectively detected in ANA ELISA I
compared to ANA ELISA II. This difference may again
be explained by the fact that for the former ELISA,
both IgM and IgG antibodies are detected, while
the latter is limited to detection of IgG antibodies.
Alternatively, there may be differences in the antigen
structures between naturally and recombinantly
formed ligands that could account for these differ-
ences. As IgM antibodies are of less clinical value [24],
the detection of this immunoglobulin class should
Screening of ANA by ELISA
probably be avoided in the next generations of the
Shield ELISA.
Nevertheless, both ELISA systems seem to fulfill
necessary qualitative requirements. This is further
strengthened by the fact that our participation in
national and international quality assessment pro-
grams has not uncovered any ANA test result deviat-
ing from the expected ones when using these two
ANA screening ELISA systems.
As both ANA screening ELISA systems and ELISA
for detection of individual ANA specificities are
expensive tests, there may be a conflict between
economy and quality when selecting a procedure for
quantifying the antibodies. The best quantitative infor-
mation is probably achieved through serum titration, a
procedure that yields information about the concen-
tration of a given antibody beyond the range that can
be measured using one serum dilution in an ELISA.
However, quantifying serum antibodies by titration is
expensive due to the need of a number of wells corre-
sponding to the number of serum dilution steps. If, as
both the current ELISA protocols recommend, one
serum dilution is used, the quantity refers to this
dilution, irrespective of whether the simple ratio
between OD of the patient serum/OD of a positive
reference control is used, or the result obtained
with patients1 serum is related to a standard curve.
The validity of this presumption is demonstrated in
Figure 4. For the two antibody examples presented
(antibodies to TPO and dsDNA), there is a very strong
correlation between the ratio (as a provisional unit,
see above) and regular units extrapolated from the
standard curves. Thus, to compromise between
economy and quality, we recommend, if a given ANA
specificity should be quantified, to use the ratio as a
semi-quantitative measure for antibody strength.
In conclusion, we have observed a modest but
similar degree of discrepancy between two classical
immunofluorescence tests for detection of ANA,
between the HEp-2 test and ANA ELISA I, and
between ANA ELISA I and ANA ELISA II. The reper-
toire of antigens used in the ELISA systems allows
detection of a dominant and clinically central popula-
tion of ANA specificities. Thus, ANA screening ELISA
systems like those studied here can in our opinion
fully substitute for IIF tests. This has the following
advantages: the tests can be fully automated if
adapted to automatic ELISA processors; the repertoire
of antigens can continuously be modified according to
future concepts and priorities; one can circumvent a
second anti-ENA antibody screening of IIF ANA-
positive sera, because the wells of the ANA ELISA
contain the central ENA antigens in addition to
DNA and histones; the ELISA results are given as
objective numbers, and not subjectively read by using
a microscope.
Acknowledgements
We thank Gerd Dagsvold, Goran Kauric, Kirsti
Hokland and Elin Albrigtsen for valuable collabor-
ation in the early phases of this work, and Dr Hans
101
Nossent for advice, support and critical comments.
We are thankful to Shield Diagnostics and Pharmacia
& Upjohn for their support during parts of this work.
References
1. Oldstone M.B. 1998. Molecular mimicry and
immune-mediated diseases. FASEB J. 12: 1255–1265
2. Haaheim L.R., Halse A.K., Kvakestad R., Stern B.,
Normann O., Jonsson R. 1996. Serum antibodies from
patients with Sjøgren’s syndrome and systemic lupus
erythematosus recognize multiple epitopes on the La
(SS-B) autoantigen resembling viral sequences. Scand.
J. Immunol. 43: 115–121
3. Moens U., Seternes O.-M., Hey A.W., Silsand Y.,
Traavik T., Johansen B., Rekvig O.P. 1995. In vivo
expression of a single viral DNA-binding protein
generates systemic lupus erythematosus-related
autoimmunity to double-stranded DNA and histones.
Pro. Natl. Acad. Sci. USA 92: 12393–12397
4. Rekvig O.P., Moens U., Sundsfjord A., Bredholt G.,
Osei A., Haaheim H., Traavik T., Arnesen E., Haga
H.-J. 1997. Experimental expression in mice and
spontaneous expression in human SLE of
polyomavirus T-antigen: a molecular bases for
induction of antibodies to DNA and eukaryotic
transcription factors. J. Clin. Invest. 99: 2045–2054
5. Dong X., Hamilton K.J., Satoh M., Wang J., Reeves
W.H. 1994. Initiation of autoimmunity to the p53
tumor suppressor protein by complexes of p53 and
SV40 large T antigen. J. Exp. Med. 179: 1243–1252
6. Klinman D.M., Steinberg A.D. 1987. Systemic
autoimmune disease arise from polyclonal B cell
activation. J. Exp. Med. 165: 1755–1760
7. Mohan C., Adams S., Stanik V., Datta S.K. 1993.
Nucleosome: a major immunogen for pathogenic
autoantibody-inducing T cells of lupus. J. Exp. Med.
177: 1367–1381
8. Desai-Mehta A., Mao C., Rajagopalan S., Robinson T.,
Datta SK. 1995. Structure and specificity of T cell
receptors expressed by potentially pathogenic
anti-DNA autoantibody-inducing T cells in human
lupus. J. Clin. Invest. 95: 531–541
9. van Venrooij W.J. 1994. Autoantigens in connective
tissue diseases. In Immunology of connective tissue
diseases. G.S. Panayi, K. Whaley, eds. Kluwer
Academic Publishers, Dordrecht pp. 305–334
10. Tan E.M. 1989. Antinuclear antibodies: diagnostic
markers for autoimmune diseases and probes for cell
biology. Adv. Immunol. 44: 93–151
11. Rekvig O.P., Andreassen K., Moens U. 1998. Anti-DNA
antibodies — towards an understanding of their origin
and pathophysiological impact. Scand. J. Rheum. 27:
1–26
12. Hess E.V., Mongey A.-B. 1992. Drug-related lupus: the
same as or different from idiopathic disease? In
Systemic lupus erythematosus. R.G. Lahita, ed. Churchill
Livingstone, New York, Edinburgh, Melbourne, Tokyo
pp. 893–904
13. Anderson P. 1977. Correlation of smooth-muscle and
nuclear antibodies in normal subjects. Clin. Exp.
Immunol. 27: 74–77
14. Tomer Y., Shoenfeld Y. 1988. Ageing and
autoantibodies. Autoimmunity 1: 141–149
102
15. Kaplan E.M., Tan E.M. 1968. Antinuclear antibodies in
infectious mononucleosis. Lancet 1: 561–563
16. Shoenfeld Y., Cohen I.R. 1987. Infection and
autoimmunity. In The antigens. Vol VII. M. Sela, ed.
Academic Press, San Diego, p 317.
17. Rekvig O.P., Hannestad K. 1977. Certain polyclonal
antinuclear antibodies crossreact with the surface
membrane of human lymphocytes and granulocytes.
Scand. J Immunol. 6: 1041–1054
18. Humbel R.L. 1993. Detection of antinuclear antibodies
by immunofluorescence. In Manual of biologic markers
of disease. Section A2. W.J. van Venrooij, R.N. Main,
eds. Kluwer Academic Publishers, Dordrecht,
pp 1–16
19. Isenberg D.A. 1994. Anti-DNA antibodies—some
enigma variations. The Immunologist 2: 190–193
E. Olaussen and O. P. Rekvig
20. Radic M.Z., Weigert M. 1994. Genetic and structural
evidence for antigen selection of anti-DNA antibodies.
Annu. Rev. Immunol. 12: 487–520
21. Marion T.N., Tillman D.M., Jou N.T., Hill R.J. 1992.
Selection of immunoglobulin variable regions in
autoimmunity to DNA. Immunol. Rev. 128: 123–149
22. Craft J.E., Hardin J.A. 1987. Linked sets of antinuclear
antibodies: what do they mean? J. Rheumatol. 14(Suppl.
13): 106–109
23. Latchman D.S. 1998. Transcription factors and cell type
specific gene expression. In Eucaryotic transcription
factors D.S. Latchman eds. Academic Press, San Diego,
pp. 119–156
24. Cohen I., Young D. 1991. Autoimmunity, microbial
immunity, and the immunological homunculus.
Immunol. Today. 12: 105–110