Anti-DFS70 antibodies in 597 healthy hospital workers




Autoantibodies against DFS70 (dense fine speckles 70) antigen (also known as lens epithelium–derived growth factor) have been recently identified among the antinuclear antibodies (ANAs) in patients with atopic disorders. We undertook this study to examine the frequency of anti-DFS70 antibodies in a large number of healthy people.


Sera of 597 healthy individuals working in a hospital (142 men, 455 women) were analyzed for ANAs and for anti-DFS70 antibodies by indirect immunofluorescence (IIF) with HEp-2 cells as a substrate and by immunoblotting using DFS70 recombinant protein and whole HeLa cell extract.


ANAs were present in 20% of all individuals by IIF. Nine percent of subjects were ANA positive at a serum dilution of 1:40, 4.0% at 1:80, 5.5% at 1:160, 1.0% at 1:320, and 0.3% at 1:640. There were 64 anti-DFS70 antibody–positive individuals. Surprisingly, this was 11% of the whole population and 54% of the ANA-positive population. The percentage of female anti-DFS70 antibody–positive subjects (86%; 55 of 64 subjects) was higher than the percentage of female anti-DFS70 antibody–negative subjects (75%; 398 of 533 subjects) (P < 0.05). The prevalence of anti-DFS70 antibody–positive sera decreased with increasing age (P = 0.0017).


Considering that anti-DFS70 antibody positivity is rare in patients with systemic autoimmune diseases, introducing the anti-DFS70 antibody examination as a screening test for ANA-positive persons could be used to rule out systemic autoimmune diseases, resulting in considerable cost-saving potential. In addition, this test defines a subpopulation of healthy people in whom long-term followup might reveal health-related implications of this finding, since anti-DFS70 antibodies have been shown to be associated with some illnesses.

Antinuclear antibodies (ANAs) are found in sera from patients with various systemic autoimmune diseases as well as in sera from healthy people (1–14). DFS70, also known as lens epithelium–derived growth factor (LEDGF), is a recently identified major ANA with a characteristic pattern in immunohistochemistry showing a nuclear distribution of dense, fine speckles (DFS) as the most common ANA pattern in patients with atopic dermatitis (AD) (15, 16). Since a 70-kd protein was recognized on immunoblotting (IB), the antigen was termed dense fine speckles 70 (DFS70).

Anti-DFS70 antibody was initially identified among ANAs from a patient with interstitial cystitis (IC) (17). When comparisons of the antibody's prevalences in various diseases were carried out by IB with the DFS70 recombinant protein, the highest prevalence was found in Japanese AD patients (30%) (15). The antibody's second-highest prevalence was in asthma patients in the US (16%), while 8.7% of IC patients in the US had the antibody. In contrast, in a limited number of healthy persons in the US or in persons with various systemic autoimmune diseases, the prevalence ranged only from 0% to 2.5%, except for 7% in the case of Sjögren's syndrome (SS) (15). On the other hand, a study investigating the prevalence of anti-LEDGF antibodies in 650 sera from a blood bank showed a prevalence of 5.4% by enzyme-linked immunosorbent assay (ELISA) only (18).

In this study, we analyzed ANAs and anti-DFS70 antibody in 597 healthy workers in an urban hospital using indirect immunofluorescence (IIF) and IB as well as an ELISA system (IgG class) that we established. Our study showed that more than half of the ANA-positive healthy individuals had anti-DFS70 antibody.


Serum samples.

Sera were obtained from 597 workers in Social Insurance Chukyo Hospital who were enrolled in our study. These workers were defined as healthy since they were employed full time and were not prevented from working because of any disabling health condition. This hospital, with 683 beds for inpatients, is located in a seaside industrial zone in Nagoya, a city with a population of ∼2.2 million in the central part of Japan. Informed consent was obtained from all volunteers. Approval was also obtained from the Ethics Committee of the hospital for this research and for the contents of the questionnaires.

The classification of occupation, sex, and mean age of the volunteers is shown in Table 1. The mean ± SD age of all volunteers was 32 ± 9 years (range 21–72 years). For a control, sera obtained from 200 Japanese patients with systemic rheumatic diseases were also examined. Patients were diagnosed as having systemic lupus erythematosus (SLE), systemic sclerosis (SSc), SS, polymyositis/dermatomyositis (PM/DM), or rheumatoid arthritis (RA) according to established criteria for these diseases (19–23). Sera from patients with SLE, SSc, SS, or PM/DM were collected from the serum bank in the Department of Dermatology at Nagoya University Hospital; some of these sera had been used in our previous studies (24, 25). Sera from RA patients were provided by one of us (Dr. Takasaki) at Juntendo University Hospital.

Table 1. Classification of volunteers
OccupationNo. of subjects (age, mean ± SD years)
  • *

    Includes 11 pharmacists, 12 x-ray technicians, 3 nutritionists, 4 physical therapists, 26 laboratory technicians, and 2 speech therapists, among others.

Nurse342 (31 ± 9)5 (32 ± 4)337 (31 ± 9)
Technical expert*126 (33 ± 9)57 (35 ± 9)69 (31 ± 9)
Doctor73 (33 ± 7)60 (34 ± 6)13 (32 ± 8)
Office worker55 (33 ± 11)22 (37 ± 14)33 (30 ± 7)
Unknown1 (31)01 (31)
Total597 (32 ± 9)144 (35 ± 9)453 (31 ± 9)

IIF microscopy.

ANAs were evaluated by a standard IIF technique with HEp-2 cells (Fluoro HEPANA Test; MBL, Nagoya, Japan) as the substrate. Sera were considered to be positive for ANAs if immunofluorescent staining was observed at a serum dilution of 1:40 according to the manufacturer's protocol. Fluorescein isothiocyanate–conjugated goat polyclonal antibody to human immunoglobulin (IgG + IgA + IgM heavy and light chain) (fluorescence to protein ratio 1.8) was used as the second antibody.

Purification of DFS70 recombinant protein.

The recombinant protein was produced with plasmid DNA (the pET 28a vector with the insert DNA coding the full open-reading frame of DFS70 [15]) transformed in competent Escherichia coli cells (BL21-CodonPlus-RIL strain; Stratagene, La Jolla, CA) according to the manufacturer's protocol. Using TALON metal affinity resins (Clontech, Palo Alto, CA), the His-tagged recombinant protein was purified according to the manufacturer's protocol.


Whole cell extracts of HeLa cells were obtained by a method described previously (26). Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transfer of proteins (cell extract or bacterial fusion protein) from 12.5% polyacrylamide gel onto polyvinylidene difluoride membrane (Millipore, Bedford, MA) were performed as described previously (27). Strips of membrane were incubated with the sample sera at 1:100 dilution. The antibody–antigen complexes were detected with horseradish peroxidase (HRP)–conjugated goat anti-human IgG (Dako, Glostrup, Denmark) at 1:1,000 dilution. Color development was carried out with Konica Immunostain (Konica, Tokyo, Japan).

ELlSA for the detection of anti-DFS70 antibodies.

ELISA was performed as described previously with slight modifications (28). Purified recombinant DFS70 protein (1 μg/well) was coated on wells of microtiter plates (ImmunoPlate MaxiSorp; Nunc, Roskilde, Denmark). E coli cells (BL21-CodonPlus-RIL strain) that did not harbor plasmid DNA were cultured similarly as described above, and the centrifuged pellet was sonicated for 20 minutes in phosphate buffered saline (PBS). After the centrifugation at 10,000g for 20 minutes, the supernatant fluids, referred to as “E coli extracts,” were collected. Next, 100-μl serum samples diluted 1:200 or serially 2-fold in reaction buffer (T-PBS [PBS containing 0.05% of Tween 20] containing 0.1 mg/ml of E coli extracts) were incubated for 60 minutes at 37°C. After the centrifugation at 10,000 revolutions per minute for 10 minutes, the absorbed sera were used as a first antibody. Anti-human IgG antibody conjugated with HRP (Dako) was used as a second antibody. The anti-DFS70 antibody level was expressed as a unit calculated from the optical density (OD) read on a representative positive sample (no. 425) chosen for reference. The ELISA unit was calculated as the value for each patient divided by that of the representative sample (no. 425).

Serum was determined to be ELISA positive when its OD was more than the mean + 2SD of 60 examples arbitrarily chosen among the samples that were negative for anti-DFS70 antibodies by IIF and IB. When the cutoff level was set at the mean + 2SD, the degree of “sensitivity multiplied by specificity” was highest (89% × 94% = 83.66%). When the cutoff level was set at the mean + 3SD, the degree of “sensitivity multiplied by specificity” was lower (84% × 96% = 80.64%).

Statistical analysis.

The chi-square test was employed for analysis of frequencies using Excel 2001 for Macintosh (Microsoft, Redmond, WA). The Mantel-extension method was used for comparing the positivities for anti-DFS70 antibodies and other ANAs between each age group. For each sex analysis, the Mantel test was used. Pearson's correlation analysis was used to compare correlations of titers between the IIF analysis and the ELISA. P values less than 0.05 were considered significant.


ANAs in IIF analysis.

ANAs were present in 119 of the 597 subjects (20%) by IIF. Fifty-four subjects (9%) were ANA positive at a serum dilution of 1:40, 24 subjects (4.0%) at 1:80, 33 subjects (5.5%) at 1:160, 6 subjects (1.0%) at 1:320, and 2 subjects (0.3%) at 1:640 (Table 2). Since some sera had concomitant patterns, the total numbers shown in Table 2 exceeded 119. Eighty-eight sera presented the speckled pattern. Among these, 69 had the DFS pattern (12% of all 597 subjects and 58% of ANA-positive subjects). The DFS pattern was unique and was characterized by staining of the chromosomes in mitotic cells as well (15, 17). It could be clearly distinguished from other homogeneous patterns or nuclear fine-speckled pattern using IIF (Figure 1A). Twelve sera had patterns concomitant with the DFS pattern: 7 as nuclear dots (like a coiled body), 3 as cytoplasmic, and 2 as a nuclear dots ring pattern. Notably, all 8 sera that were ANA positive at high titer (≥1:320) had the DFS pattern. One of these 8 sera had a concomitant nuclear dots pattern. Of ANAs other than those with the speckled pattern, 11 had the homogeneous pattern, 4 had the nucleolar pattern, 1 had the discrete speckled pattern, 15 had the cytoplasmic pattern, and 16 had other patterns.

Table 2. Indirect immunofluorescence patterns and serum titers*
  • *

    The total numbers at each titer exceeded the number described in the text because several sera had concomitant patterns.

 Dense, fine-speckled2311276269
 Other speckled11620019
Discrete speckled010001
Figure 1.

A, Indirect immunofluorescence (IIF) of anti–dense fine speckles 70 (anti-DFS70) antibody–positive sera using HEp-2 cells as substrate. The DFS pattern was unique and characterized by staining of chromosomes in mitotic cells as well (arrows). B, Immunoblotting (IB) using DFS70 recombinant protein to show the presence of anti-DFS70 antibodies. C, IB using whole HeLa cell extract to confirm the specificity of the sera reacting to the recombinant protein. In B and C, lanes 1 and 2 show anti-DFS70 antibody–positive sera; lane 3 shows anti-DFS70 antibody–negative sera.

Detection of anti-DFS70 antibody by IB and grouping of samples into 3 categories.

Since DFS sera were detected in the IIF analysis much more than we had expected, we next investigated the presence of anti-DFS70 antibodies in all volunteers' sera using the IB analysis with DFS70 recombinant protein (Figure 1B). The results showed that sera from 71 of the 597 subjects (12%) were positive. In order to confirm the specificity of the sera reacting to the recombinant protein, IB with HeLa cell extract was then performed (Figure 1C). Among the 71 sera, 69 reacted with protein of ∼70 kd. Since 5 of these 69 sera were ANA negative by IIF, 64 sera were judged to be definitely anti-DFS70 antibody positive according to the results of the 3 analyses. This anti-DFS70 antibody–positive group constituted 11% of all 597 subjects and 54% of the 119 ANA-positive subjects. The 3 sera that were negative by IB using anti-human IgG antibodies showed the DFS pattern in IIF using polyclonal anti-human immunoglobulin antibody containing all IgG, IgA, and IgM classes used as the second antibody. For these sera, detection of anti-DFS70 antibodies of the IgM and IgA classes was tried by using IB, but no serum was found to have either of the IgM- or IgA-class anti-DFS70 antibodies.

Serum samples from the 597 persons analyzed were classified into the following 3 groups. Group A consisted of 64 samples which presented the DFS pattern by IIF (ANA positive) and which were anti-DFS70 antibody positive by IB. Group B consisted of 55 samples that were ANA positive but anti-DFS70 antibody negative. Group C consisted of 478 samples that were ANA negative. Table 3 shows the classifications of occupation, age, and sex in each group of serum sample donors. The percentage of women in group A (86%; 55 of 64 subjects) was higher than that in groups B and C combined (75%; 398 of 533 subjects) (P < 0.05) and was also higher than that in group C alone (74%; 354 of 478 subjects) (P < 0.04). For all occupation groups, the mean ± SD ages of men and women in group A were lower than those in groups B and C. Based on this result, we further compared the 3 groups from the aspect of the different age subgroups (Figure 2). The prevalence of anti-DFS70 antibody–positive sera decreased with increasing age (P = 0.0017 by Mantel-extension test) and, to the same extent, in women (P = 0.0066 by Mantel test). On the other hand, there was a trend toward increasing ANA positivity with age in group B (P = 0.0198 by Mantel-extension test). There were no significant differences in anti-DFS70 antibody positivity among the different occupations (data not shown).

Table 3. Serum donors by classification of occupation, age, and sex in the 3 categories of serum*
OccupationMales, no. (age, mean ± SD years)/females, no. (age, mean ± SD years)
Group AGroup BGroup C
  • *

    Group A = sera presenting the dense, fine-speckled (DFS) pattern by indirect immunofluorescence (antinuclear antibody [ANA] positive) and anti-DFS70 antibody positive by immunoblotting; group B = ANA-positive, anti-DFS70 antibody–negative sera; group C = ANA-negative sera.

Nurse1 (27)/42 (27 ± 6)0/33 (33 ± 11)4 (33 ± 4)/262 (31 ± 9)
Technical expert2 (26 ± 2)/9 (29 ± 7)3 (44 ± 9)/7 (31 ± 9)52 (35 ± 9)/53 (31 ± 9)
Doctor5 (33 ± 5)/1 (30)3 (36 ± 7)/1 (26)52 (34 ± 6)/11 (32 ± 8)
Office worker1 (25)/3 (28 ± 4)5 (45 ± 17)/3 (33 ± 11)16 (35 ± 12)/27 (31 ± 7)
Unknown0/00/00/1 (31)
Total9 (30 ± 5)/55 (28 ± 6)11 (43 ± 13)/44 (33 ± 10)124 (34 ± 8)/354 (31 ± 9)
Figure 2.

Prevalence of 3 types of sera according to age group and sex of donors. Group A (solid sections of bars) consisted of 64 serum samples which presented the DFS pattern by IIF (antinuclear antibody [ANA]–positive sera) and which were anti-DFS70 antibody positive by IB. Group B (shaded sections of bars) consisted of 55 samples that were ANA positive but anti-DFS70 antibody negative. Group C (open sections of bars) consisted of 478 ANA-negative samples. Values in bars are percentages of each type of serum within each bar. See Figure 1 for other definitions.

Measurement of anti-DFS70 antibodies by the ELISA system.

We applied the ELISA system to measure anti-DFS70 antibodies in 597 samples. As a result, 82 sera were judged to be positive by ELISA and 515 to be negative by ELISA. Among the 64 sera of group A, 57 were true positive and 7 were false negative by ELISA (sensitivity 89%). In groups B and C combined (a total of 533 sera), 508 sera were true negative and 25 were false positive (specificity 95%) (Figure 3). In group A, the titers showed a significant correlation (P < 0.001) between ELISA and IIF, as shown in Figure 4.

Figure 3.

Enzyme-linked immunosorbent assay (ELISA) analysis of anti–dense fine speckles 70 (anti-DFS70) antibody levels in sera from 597 healthy hospital workers. The anti-DFS70 antibody level was expressed as a unit calculated from the optical density read on a representative positive sample chosen for reference. Values shown (ELISA units) were calculated as the value for each patient divided by that of the representative sample. The dashed line indicates the cutoff level (0.298 ELISA units; mean + 2SD of 60 anti-DFS70–negative sera). Each large circle represents 50 persons. Each small circle represents 1 person. NG = negative group; PG = positive group.

Figure 4.

Correlations between anti-DFS70 antibody titers obtained by enzyme-linked immunosorbent assay (ELISA) and titers of antinuclear antibodies (ANAs) presenting the DFS pattern by IIF in anti-DFS70 antibody–positive sera. The correlation between the IIF titers (x-axis) and the ELISA titers (y-axis) was significant (r = 0.556, P < 0.001). The slope of the correlation line was expressed as y = 0.46 + 0.24x, with x = 1 at a serum dilution of 1:40, x = 2 at a serum dilution of 1:80, x = 3 at a serum dilution of 1:160, x = 4 at a serum dilution of 1:320, and x = 5 at a serum dilution of 1:640. See Figure 1 for other definitions.

Investigation of the health condition using the questionnaires.

Questionnaires on present and past health conditions were given to all subjects in groups A and B and to 100 subjects in group C. These 100 subjects were selected arbitrarily from among those in group C whose age and sex were matched with those of subjects in groups A and B. We collected replies from 52 of 64 persons in group A, 45 of 55 in group B, and 68 of 100 in group C (75% of valid response rates).

Analyses of the questionnaires indicated that ≥10% of the respondents had diseases such as AD (group A, 21%; group B, 18%; group C, 12%) and allergic rhinitis (group A, 35%; group B, 36%; group C, 32%), but there were no significant differences between the groups. None of the subjects had either rheumatic disease or connective tissue disease. Few of the subjects had illnesses such as heart, lung, kidney, liver, stomach, neurologic, or genital disease. For the questions related to the symptoms of chronic fatigue syndrome (CFS) (29), although none of the subjects fulfilled the CFS criteria, there were many replies, as follows: “postexertional malaise lasting more than 24 hours” (group A, 27%; group B, 9%; group C, 10%), “new headaches” (group A, 23%; group B, 20%; group C, 10%), “impaired memory” (group A, 17%; group B, 20%; group C, 24%), “impaired concentration” (group A, 27%; group B, 22%; group C, 26%), and “depression” (group A, 23%; group B, 13%; group C, 10%). In response to the question concerning “postexertional malaise lasting more than 24 hours,” group A showed higher frequencies than group B (P < 0.03), group C (P < 0.02), or groups B and C combined (P < 0.01). Concerning the incidence of diarrhea, there was a significantly higher frequency in group A (19%) than in group C (6%) (P < 0.03) or in groups B (9%) and C combined (P < 0.03), although this item was not among the CFS criteria.

Detection of anti-DFS70 antibody in 200 patients with systemic rheumatic disease.

We checked for anti-DFS70 antibody in 200 Japanese patients with different systemic rheumatic diseases using both IIF and IB methods. The numbers and sex distributions of each disease group are shown in Table 4. We found 3 anti-DFS70 antibody–positive sera (3 of 200 sera = 1.5% positivity). These findings were similar to those obtained in a US population (15). It is not possible to simply compare the frequencies between healthy individuals and patients with rheumatic disease, because these two groups were not matched by age and sex. However, the 3 anti-DFS70 antibody–positive cases were not restricted to young people (they occurred in persons age 49, 51, and 53 years). Moreover, when we analyzed the data from persons under age 50, the frequency of the antibody was still significantly higher in healthy individuals than in the patients with rheumatic disease (64 of 564 healthy individuals versus 1 of 101 patients; P < 0.002 by Fisher's exact test). Therefore, this antibody is not very common in Japanese, and it also appears that positivity for this antibody is very infrequent among patients with systemic rheumatic disease.

Table 4. Anti–dense fine speckles 70 (anti-DFS70) antibodies in 200 Japanese patients with systemic rheumatic disease*
DiseaseNo. of patients examined (age, mean ± SD years)No. of patients with anti-DFS70 antibodies (no. male/no. female)Frequency, %
  • *

    SLE = systemic lupus erythematosus; SSc = systemic sclerosis; RA = rheumatoid arthritis; SS = Sjögren's syndrome; PM/DM = polymyositis/dermatomyositis.

SLE55 (42 ± 12)1 (20)/54 (42 ± 12)1 (0/1)2
SSc50 (52 ± 10)4 (49 ± 8)/46 (52 ± 10)00
RA40 (51 ± 12)0 (56 ± 9)/30 (49 ± 12)00
SS30 (47 ± 12)3 (51 ± 17)/27 (47 ± 11)2 (0/2)7
PM/DM25 (48 ± 14)9 (44 ± 19)/16 (51 ± 9)00


IIF for detection of ANAs is a very important tool for the screening of autoimmune diseases. However, it is well known that ANAs found to be positive by IIF are sometimes present in healthy people, and that up to 30% of healthy adults are ANA positive when HEp-2 cells are used as substrates (1–14). In our study, 119 of 597 volunteers (20%) were found to have ANAs. Very surprisingly, more than half of the ANA-positive individuals, which corresponded to ∼11% of the healthy workers involved in the study, had anti-DFS70 antibodies. The “healthy people” were variously defined in previous reports, such as residents in certain areas (12), blood donors (11), people in nursing homes (8, 9), and those who were considered to be healthy by the results of questionnaires and/or physical examinations (8, 9, 12). In this study, we adapted the definition of “healthy people” according to a previous report (13): individuals who were working full time on a regular and daily basis and who did not have any physical or mental disabilities, among whom some could have been taking common medications that did not affect or interfere with their daily work activities.

Although anti–single-stranded DNA (anti-ssDNA) antibodies, etc., were previously reported in healthy people, the rate of positivity was very low except in a few reports (8–10). In the present study, we tried to detect various autoantibodies with clinical “markers” including anti-Sm, anti-RNP, anti–Scl-70, anti–Jo-1, anti-SSA, anti-SSB, anti–CENP-B, anti–double-stranded DNA (anti-dsDNA), and anti-ssDNA in all sera using commercial ELISA kits (MBL) and/or IB with HeLa cell extract; we also tried to detect anti-dsDNA antibodies using IIF with Crithidia luciliae as a substrate (data not shown). Among these, anti-ssDNA antibodies were found in 11 of the 597 sera (2%). Anti-SSA and anti–CENP-B were present in only a few sera (6 of 597 [1%] and 1 of 597 [0.2%], respectively), and the others were not. Therefore, anti-DFS70 antibody was the most frequently identified ANA in the healthy people.

After the DFS70 complementary DNA which we cloned from IC patients was deposited in GenBank (accession no. U94319) in 1997, two groups independently reported clones identical to DFS70. One is called “transcription coactivator p75” (30), and the other is LEDGF (31). DFS70/LEDGF is a very interesting molecule which has many biologically significant roles. LEDGF is a factor that enhances the survival of various cell types under stress. LEDGF is also a transcriptional activator, and it binds to promoter elements of heat-shock and stress-related genes to activate the expression of these genes. The elevated levels of the stress-related family of proteins, such as heat-shock proteins, antioxidant proteins, and detoxication enzymes, might suppress apoptosis induced by stress.

Since our established ELISA system (IgG class) for detecting anti-DFS70 antibody has high sensitivity and specificity, it should be suitable for the mass screening of persons for anti-DFS70 antibody. Moreover, judging from the good correlation between the ELISA titers and the IIF titers, this system can be also utilized for quantitative assays. In a study by Ayaki et al (18), the prevalence of the autoantibody to LEDGF in sera from 650 donors was 5.4% by ELISA only, and there was no significant difference in anti-LEDGF reactivity among the age groups. The reason for the differences from our findings may come from the definition of antibody positivity or the sample population. However, two caveats should be noted about the study by Ayaki et al: 1) since sera were obtained from blood bank donors, these samples were not well characterized; and 2) Ayaki et al did not discuss the relationship between LEDGF and ANA.

Since there were discrepancies in the results among the IIF, the IB, and the ELISA, immunoprecipitation assays with the in vitro transcription product of DFS70 were further performed to check the results, and the grouping in the present study was confirmed (data not shown). These discrepancies were probably caused by differences in the nature of the antigens (denatured, native) or in the sources of the antigens (cellular, recombinant) among the IIF, the IB, and the ELISA. We would like to stress that such discrepancies occurred in only a very small population in the present study.

The notable point is that anti-DFS70 antibodies were found significantly more often in young people under age 35 than in people over age 35 (P < 0.003) (Figure 2). In a previous report on the prevalence of ANAs in normal individuals, no differences were observed among 4 age subgroups (21–30, 31–40, 41–50, and 51–60 years) (13). We also found no significant differences in the rate of ANA positivity (group A plus group B) among the age subgroups (P = 0.458 by Mantel-extension test) (Figure 2).

We reported previously that anti-DFS70 antibodies were found more often in AD patients ages 4–43 years (15). When the data from the AD patients over age 20 in the previous study were age- and sex-matched with the data from the subjects in the present study, anti-DFS70 antibodies were still more frequently found in AD patients than in the healthy people of the present study (11 of 48 versus 6 of 48), but the difference was not statistically significant. Although the results of the questionnaires on the subjects' past and present health conditions showed a high prevalence of AD in group A, anti-DFS70 antibody–positive people were not so numerous compared with the other groups. Similar frequencies among the 3 groups were observed in other allergy disorders (data not shown). The incidences of AD (in the age range 5–18 years) and allergic rhinitis in Japan are 11–24% (32) and 30% (33), respectively, and similar rates were obtained in the present study. Recently, investigators in our group reported that anti-DFS70 antibodies were also found in some patients with CFS, and that these antibodies were thought to be the major ANA in Japanese CFS patients (34). It is of interest that the anti-DFS70 antibody–positive subjects among the healthy individuals had some possible symptoms of CFS such as “postexertional malaise,” “depression,” etc.

It is known that autoreactive antibodies and autoreactive T cells are present even in healthy individuals and in virtually all vertebrate species (35). Antibodies that are present in the serum of healthy individuals in the absence of deliberate immunization with any antigen are referred to as natural autoantibodies (35). Our data suggest that the anti-DFS70 antibodies could possibly be defined as natural autoantibodies. It has been reported that a homuncular self antigen is expressed in the thymus and is recognized by natural autoantibodies and T cells in healthy subjects (36). Interestingly, the messenger RNA of transcription coactivator p75, which corresponds to DFS70, is expressed more in the thymus than in other tissues (30). Therefore, DFS70 in the thymus might be recognized as a homuncular self antigen. The lower frequency at older ages might be associated with the involution of the thymus. It would therefore be interesting to survey anti-DFS70 antibodies in healthy people under age 20. Sex differences in the production of the antibody were also interesting. Sex hormones might influence the aberrant expression of the antigen, which could lead to the production of the autoantibody.

In the case of thyroglobulin as an autoantigen, the autoantibodies from healthy individuals recognize epitopic clusters that are clearly distinct from those recognized by autoantibodies from patients with Hashimoto thyroiditis (37). In future studies of the epitope mapping of DFS70, the differences between the anti-DFS70 antibodies associated with disease and natural autoantibodies will be clarified. It would also be important to study the antibody differences between patients and healthy people in terms of titers, immunoglobulin subclass, idiotype antibodies, etc. Previous reports showed high frequencies of antibodies in AD (18 of 64 patients, 28%) (15) and in CFS (18 of 115 patients, 16%) (34). Changes that occur in the titers of the antibodies, as well as diseases that develop in the future, should be observed in anti-DFS70 antibody–positive persons, especially those with high titers of the antibodies.

In the early 1980s, sections of mouse livers and rat kidneys as tissue substrates for the ANA test were replaced by cultured human laryngeal epithelial carcinoma (HEp-2) cells. Because of the high sensitivity of HEp-2 cells to small amounts of ANAs presenting in patients and controls, false-positive samples still exist at present. According to our clinical experience, sera from young women often showed the speckled or speckled–homogeneous patterns in the IIF test, but these subjects were not always diagnosed as having autoimmune diseases such as lupus after long-term observations. The finding of anti-DFS70 antibodies in such people would rule out systemic autoimmune diseases, resulting in considerable cost-saving potential. Until the present study, the immunologic specificity or target antigens of these ANAs have largely escaped identification. We now show that more than one-half of such antibodies are anti-DFS70 antibodies which are not specific for any one disease condition, but which are more prevalent in AD, asthma, and IC, and perhaps in other as-yet-undetermined disease conditions. Further, similar studies from various countries will be needed to investigate possible racial differences in the prevalence of this antibody.


The authors thank Dr. Kenji Wakai at the Department of Preventive Medicine, Nagoya University Graduate School of Medicine, for analyzing the Mantel tests.