å¡ ORIGINAL
ARTICLE
å¡
Incidence of Anti-Mouse IgG in Normal Subjects and Patients with Autoimmune Thyroid Disease Makoto Iitaka, Shinji Kitahama, Nobuhiko Fukasawa, Shiro Miura, Yoshito Kawakami, Shinichi Sakurai, Morifumi Yanagisawa and Jun Ishii
Serum antibodies to mouse IgG occasionally interfere with two-site immunometric assays in which mouse monoclonal antibodies are used. We examined the titers of antibodies to mouse IgG in serum samples from normal subjects and patients with autoimmune thyroid disease (AITD) using an enzyme-linked immunosorbent assay. Anti-mouse IgG antibodies were positive in 7/119 patients with Graves' disease (5.9%) and 3/60 patients with Hashimoto's thyroiditis (5.0%). One of the 15 patients with a thyroid neoplasm (6.7%) also had antibodies to mouse IgG, as did 5/60 healthy subjects (8.3%). These antibodies were either of the IgG or IgM class. There was no significant difference in the incidence of positive anti-mouse IgG antibody between normal subjects and patients with AITD. It is important to note this high incidence of antibodies to mouse IgG due to the potential of interference with immunometric assays employing mouse monoclonal (Internal Medicine 31: 984-988, 1992) antibodies. Key words: heterophilic antibody, two-site immunometric assay
Introduction
for anti-mouse IgG antibodies. We found that a con siderable number of patients with AITD had antibodies to mouse IgG. However, there was no significant difference in the incidence of antibodies to mouse IgG between the AITD patients and the control subjects.
Heterophilic antibodies are known to interfere with numerous of important immunoassays, such as those for or-fetoprotein (1), hepatitis B surface antigen (2), ferritin (3), hCG (4), and TSH (5-9). A 7% incidence of antibodies Materials and Methods against ovine or bovine immunoglobulin has been reported in healthy individuals (1). Antibodies to the immuno Su bjects globulins of various animal species, such as rabbit (3, 5-9), One-hundred and nineteen patients with Graves' disease guinea pig (2), and goat (4) have also been found. The recent (21 men and 98 women) and 60 patients with Hashimoto's development of two-site immunometric assays using mouse thyroiditis (8 men and 52 women) were examined. As monoclonal antibodies has enabled higher specificities and controls, 15 patients with thyroid tumors (10 with benign sensitivities. However, serum antibodies to mouse IgG may adenoma and 5 with cancer, 2 men and 13 women), and interfere with many of the assay kits employing mouse 60 apparently healthy subjects (18 men and 42 women) who monoclonal antibodies (10-16). To prevent such inter did not have antibodies to thyroglobulin or thyroid ference, nonimmune mouse serum or IgG are commonly microsomes were also studied. The serum was separated included in these kits. However, we previously encountered immediately after the blood withdrawal and stored at a patient with Graves' disease whose serum interferred with -20°C. the TSH assay kit that included mouse IgG (17). This Measurement of anti-mouse IgG antibodies prompted us to examine the incidence of serum antibodies The titers of antibodies to mouse IgG were measured to mouse IgG in patients with autoimmune thyroid disease using an enzyme-linked immunosorbent assay (ELISA). One (AITD). As "nonimmune" controls, serum from healthy From the Fourth Department Internal Medicine, Saitama individuals and patients withof thyroid tumors was also testedMedical School, Saitama Received for publication December 4, 1991; Accepted for publication May 7, 1992 Reprint requests should be addressed to Dr. Makoto Iitaka, the Fourth Department of Internal Medicine, Saitama Medical School, 38 Mo Moroyama, Iruma-gun, Saitama 350-04, Japan 984
Internal
Medicine
Vol. 31,
No. 8 (August
1992)
Incidence
of Anti-Mouse IgG Antibody
milligram of purified mouse IgG (Sigma Chemical Co. , St. Louis, MO) was labelled with 60 ^g of N-hydroxysuccini midobiotin (Pierce, Rock ford, IL). Ninety-six well plates (Immunoplate, Nunc, Denmark) were coated with 10 mg/1 of purified mouse IgG in 0. 1 mol/1 carbonate-bicarbonate buffer, pH 8.3. Plates were then incubated with 1 g/1 bovine serum albumin (Sigma) in phosphate-buffered saline (PBS) to reduce nonspecific binding. Serum samples (50/ul) were incubated in the plates for 1 hour at room temperature. The plates were then washed 3 times with PBS containing 0. 1 % Tween 20 (washing buffer), biotinylated mouse IgG was added, and the plates were incubated at room temperature for 1 hour. After washing 3 times with washing buffer, the plates were incubated with avidin-conjugated horseradish peroxidase (Sigma) at room temperature for 1 hour. The plates were then washed an additional 3 times with washing buffer, and 0.4 g/1 of o-phenylenediamine (Wako Chemical Co., Osaka, Japan) in 0.1 mol/1 citrate buffer, pH 4.8, con taining 0.003% hydrogenperoxide was added. After 30 minutes of incubation at room temperature, 50#1 of 6 mol/1 H2SO4 was added to stop the reaction. The optical density (O.D.) at 492 nm was then measured using an ELISA reader. Affinity-purified goat (IgG) antibodies to mouse IgG (Cappel, West Chester, PA) were diluted to 1:10,000 and used as a standard. Titers for the anti-mouse IgG antibodies were expressed as the ratio of the O.D. of the serum sample to that of the standard (ELISA index). To examine the
solution containing 250 mg/1 of 3,3'diaminobenzidine (Wako) and 0.01 % hydrogenperoxide in 0.1 mol/1 Tris HC1 buffer, pH 7.4. Immunoglobulins bound to biotinylated mouse IgG were detected as a brown precipitation line. Laboratory measurements Serum thyroid hormone and TSH levels were measured using commercially available kits (Amerlex-M, Amersham International Pic, Buckinghamshire, England; and RIA gnost hTSH, Behringwerke, Marburg, FRG, respectively). The normal ranges of serum free thyroxine (FT4) and TSH were 9.7-24.5 pmol/1 and 0.3-3.5 mU/1, respectively. Antibodies to thyroglobulin (TGPA) and thyroid micro somes (MCPA) were detected using the kits obtained from Fuji Rebio (Tokyo, Japan). TSH receptor antibodies (TRAb) were determined by a radioreceptor assay (S.R.S. Ltd, Cardiff, UK). Statistical analysis The Smirnov-Grubbs test was used to exclude extremely deviated data in the determination of the normal range. The chi-squared test was used to compare the incidence of serum antibodies to mouse IgG in the healthy subjects and in the patients with thyroid disease. Results
Specificity and accuracy of the ELISA for anti-mouse IgG an tibodies To evaluate the specificity of this assay, bovine, rabbit, and mouse IgG were added to serum samples from four specificity of the assays, varying amounts of rabbit, bovine, subjects positive for antibodies to mouse IgG. As shown Immunoelectrophoresis and enzymatic detection of anti in Table 1, up to 500fig of bovine and rabbit IgG and no and mouse IgG (Sigma) were added to serum samples mouse IgG antibodies effect on the assay. On the other hand, 50^g of mouse IgG obtained from the subjects for anti-mouse IgG Immunoelectrophoresis waspositive performed as previously significantly reduced the readings. The intra-assay co antibodies(17, and 18) the using ELISAgoat wasanti-human repeated. whole serum, anti described efficient of variation was 1.6-ll.8% and the interassay human IgG or anti-human IgM (Cappel). Biotinylated variation was 3.7-24.1%. Of the 60 serum samples ob mouse IgG was then applied to the plates. After overnight tained from healthy subjects, 1 gave an extraordinarily high incubation, the plates were washed extensively with washing ELISA index. This was excluded by the Smirnov-Grubbs buffer and incubated with avidin-conjugated horseradish test when determining the normal range. The mean + 2 SD peroxidase (Sigma) for 2 hours at room temperature. After of the remaining 59 samples was considered as the upper further washing, the plates were incubated with the substrate limitonofthe normal thisAnti-Mouse study. Samples Table 1. Inhibitory Effect of Various IgG Assayinfor IgG with an ELISA index Antibodies A dd ition of ( ｫg) S ubject
R abbit IgG 100 250 500
100
B ovin e IgG 250 500
M o use IgG 50
C ontro l
0 .583
0 .617
0 .479
0.500
0.625
0.502
0.0 18
0 .521
0 .354 0 .646
0 .429 0 .667
0 .354 0 .546
0.542 0.646
0.4 17 0.583
0.407 0.576
0.023 0.024
0 .4 72 0 .597
0 .621
0 .633
0 .52 1
0.59 8
0.5 12
0.503
0.02 8
0 .576
The ELISA index for anti-mouse IgG antibody was determined in the presence or absence (control) of IgG from various animal species. Internal
Medicine
Vol. 31,
No. 8 (August
1992)
985
Iitaka Table 2.
Incidence
et al
of Anti-Mouse IgG Antibodies
T o ta l
E L IS A in d e x
N u m b er o f
In cid e n c e o f
S u b jects
n um ber
(m e a n ｱ S D )
p o sitiv e p a tien ts
p o sitiv ity (% )
G D
1 19
0 .2 16 + 0 .13 7
7
5 .9
H T T N
60 15 60
0 .2 10 + 0 .12 1 0 .2 2 9 + 0 .14 0 0 .17 8 + 0 .1 16
3 1 5
5 .0 6 .7 8 .3
GD: Graves' disease, HT: Hashimoto's thyroiditis, TN: thyroid neoplasm, C: healthy control. *An ELISA index above the mean + 2 SD for 59 healthy subjects was defined as positive for anti-mouse IgG antibody, [one extremely high ELISA index value determined in a healthy subject was excluded by the Smirnov-Grubbs test.] Table 3. Clinical Profile of the Patients with Autoimmune Thyroid Disease Who were Positive for Anti-Mouse IgG Antibody P atient D x
E L ISA ind ex F T 4
TSH
TR A b
N o. 1 G D 2 GD 3 HT
0 .535 1.405 0 .660
15.3 8.5 14.2
< 0 .1 < 0.1 0.26
53 .8 55 .8 4 .5
4 5 6 7
0 .428 0 .432 0 .447 0 .590 0 .9 11 0 .569 0 .4 16
6.2 < 0 .1 13.1 1.2 7.3 0.85 20.7 2 .9 15.7 1.3 21.6 2.6 14.2 2 .7
30 .3 - 1.6 - 67.6 6 .5 5 .0 5 .7 0 .0
9 10
G G G G H G H
D D D D T ,R A D T
A TA
AM A
Ig class#
6 ,400 25 ,600 6 ,400
IgG IgG IgM
1,600 400 1,600 1,600 6 ,400 6 ,400 102 ,400 102 ,400 2 5,600 400 6 ,400 6 ,400
ND ND ND IgG IgG ND ND
100 400
Dx: diagnosis, GD, Graves' disease, HT: Hashimoto's thyroiditis, TRAb: TSH receptor antibody, ATA: anti-thyroglobulin antibody, AMA: anti-microsomal antibody, ND: notclass determined. #Immunoglobulin of anti-mouse IgG antibody. *TSH was measured with the RIAgnost TSH kit. Normal range: FT4: 9.7-24.5 pmol/1, TSH: 0.3-3.5 mU/1 TRAb: -10 to +10% above this value (>0.409) were defined as positive for anti-mouse IgG antibodies. A nti-mouse IgG antibodies in patients with thyroid disease and in healthy subjects Seven out of the 1 19 patients with Graves' disease (5.9%) had anti-mouse IgG antibodies. Three of the 60 patients with Hashimoto's thyroiditis (5.0%) and 1 of the 15 patients with thyroid neoplasms (6.7%) were also positive for these anti bodies. It is of interest that 5 of the 60 healthy subjects (8.3%) were also positive for anti-mouse IgG antibodies. As shown in Table 2, there was no significant difference in the mean ELISA index and the incidence of these anti bodies between the patients and the normal subjects. Clinical details of the patients with AITD who were positive for anti-mouse IgG antibodies are shown in Table 3. There was no correlation between the ELISA index for the anti-mouse IgG antibodies and thyroid autoantibodies or serum thyroid hormone levels in the patients with AITD. Furthermore, there was no prominent discrepancy between serum TSH and free T4 (FT4) levels, suggesting that the anti-mouse 986
antibodies in these patients did not interfere with the TSH assay used in this study. It is of interest that Graves' disease patient No. 6 had an abnormally negative value for TRAb due to the presence of serum anti-bovine TSH antibodies. Examination of longitudinally collected serum samples from two patients with Graves' disease showed that the anti mouse IgG antibodies were positive throughout the course irrespective of the serum levels of thyroid hormone or autoantibodies, although the ELISA index decreased with time in both cases (Table 4). Immunoglobulin class of anti-mouse IgG antibodies Immunoelectrophoresis and subsequent enzymatic de tection showed that antibodies to mouse IgG were either ofthe IgG or IgM class (Fig. 1). As shown in Table 3, 4 patients with AITD had anti-mouse antibodies of the IgG class. One patient with Hashimoto's thyroiditis had anti mouse IgG antibodies of the IgM class.
Internal
Medicine
Vol. 31, No. 8 (August
1992)
Incidence
of Anti-Mouse IgG Antibody
Table 4. Longitudinal Changes of the Titers of Anti-Mouse IgG suggesting that the nonimmune mouse serum or IgG in the Antibodies, the Titers of Thyroid Autoantibodies, and the kits used in this study effectively inhibited interference by Serum Thyroid Hormone Levels in the Patients with Graves' the serum antibodies. Previous studies (18-20) have shown Disease that antibodies to mouse IgG interfere with some, but not P a tie n t D a te (Y / M / D ) E L IS A in d e x FT 4 TR A b A TA A M A all, immunometric assays. This may be due to differences in the immunoglobulin class or subclass of the mouse M M 851225 0 .9 3 3 5 .3 N A 100 6 ,4 00 860605 0 .8 8 9 1 7 .9 N A 1 ,6 0 0 monoclonal antibodies in the kits or to differences in the 870525 0 .8 3 7 1 2 0 .3 4 1 .9 1 ,6 0 0 absorbing capacity of the nonimmune mouse serum or IgG 880609 0 .7 2 6 1 5 .7 3 9 .8 6 ,4 0 0 in the kits. Indeed, antibodies to mouse IgG found in our 890729 0 .6 9 3 1 4 .5 3 1 .2 1 ,6 0 0 previously reported patient did not interfere with the TSH 900602 0 .6 0 3 17 .6 1 8 .7 1 ,6 0 0 assay kit used in this study (17). These observations indicate SS 840605 0 .7 2 0 1 3 .O N A 1 0 2 ,4 0 0 that anti-mouse IgG antibodies detected in this study may 8 5 0 2 15 0 .6 7 7 1 7 .6 N A 2 5 ,6 0 0 have the potential to interfere with some other 860225 0 .6 1 3 2 9 .2 N A 400 2 5 ,6 0 0 Antibodies to mouse IgG were consistently detected in 870525 0 .4 3 1 1 2 .9 13 .2 10 0 4 0 9 ,6 0 0 immunometric assays. 880527 0 .5 1 6 1 5 .4 3 .1 10 0 6 ,4 0 0 some Graves' patients throughout the course of their 890203 0 .4 0 6 1 5 .6 6 .7 10 0 1 0 2 ,4 0 0 disease, regardless of their thyroid function or thyroid 900323 0 .4 4 7 19 .3 4 .2 2 5 ,6 0 0 autoantibody titers. There was thus no relation between TRAb : TSH receptor antibody, ATA: anti-thyroglobulin antibody, AMA: for anti-mouse IgG antibodies and thyroid auto positivity anti-thyroid microsomal antibody, NA: not available. antibodies. Furthermore, antibodies against mouse IgG were detected almost equally in the healthy subjects and the patients with thyroid neoplasms as in the patients with AITD.These observations suggest that the production of antibodies to mouse IgG is not related to thyroid auto immunity. It is of interest that one patient with Graves' disease (patient No. 6 in Table 3) had antibodies to both mouse IgG and bovine TSH. Antibodies to bovine TSH interfere with the TSH receptor antibody assay and cause extremely negative TRAb values (21). It is interesting to speculate about thistopatient develop such The incidence ofwhether antibodies sheep is orprone bovinetoimmuno heterophilic antibodies. globulin has been reported to be 7^o in a population of blood donors (1). Hedenborg et al (5) have suggested that at least 5% of all sera tested were positive for antibodies to rabbit immunoglobulin. The incidence of previously reported heterophilic antibodies thus appears to be quite Fig. 1. Immunoglobulin class of antibodies to mouse IgG. Immuno similar to that for anti-mouse IgG antibodies shown in this globulins were precipitated with antibodies to human whole serum. Immunoglobulins bound to biotin-labelled mouse IgG could be seen after study (5.0-8.3%). Our results are also compatible with the the enzymatic reaction with avidin-peroxidase and the substrate. study Px andby Thompson et al (13), who reported that about 9% P2 (patients No. 1 and 2 in Table 3) show the presence of antibodies of of blood the IgG class directed against mouse IgG. P3 (patient No. 3 in Table 3) donors had spuriously high CK-MB isoenzyme indicates the presence of an antibody of the IgM class. N is a normal values which were greatly depressed when re-assayed in the control. presence of nonimmune mouse serum. It may seem sur prising that the incidence of heterophilic antibodies is fairly Discussion high, although the incidence of interference with an assay We theantiserum immunoglobulin classto ofbe theonly anti thatalso usesexamined guinea pig was reported We previously encountered a patient with Graves' disease 0.52% (3). mouse IgG antibodies. To our knowledge, this is the first whose serum interfered with the TSH assay in which mouse report that demonstrated the presence of IgG and IgM IgG was included (17). Similar to our case, other patients classes of anti-mouse IgG antibodies. No significant who escaped from such blockade have also been reported difference was found in thyroid function tests or thyroid (18-20). In the present study, we showed that 5-6% of autoantibody titers between patients with IgG class anti bodies and those with IgM class antibodies. It remains to patients with AITD had antibodies to mouse IgG. However, be clarified whether or not there is class switching of these the patients with anti-mouse IgG antibodies had serum TSH antibodies. and thyroid hormone levels compatible with clinical status , Internal
Medicine
Vol. 31, No. 8 (August
1992)
987
Iitaka
et al
In the present study, we demonstrated that antibodies 9) Howanitz PJ, Howanitz JH, Lamberson HV, Ennis KM. Incidence and mechanism of spurious increase in serum thyrotropin. Gin Chem against mouse IgG were present in 6-8% of both patients 28: 427, 1982. with thyroid disease and healthy subjects. Although the 10) Bock JL, Furgiuele J, Segen JC. Choriogonadotropin measured with the Tandem-E immunoenzymetric assay system. Clin Chem 31:441 , precise mechanism of the production of these anti-mouse 1985. IgG antibodies remains to be clarified, it is important to 1 1) Bock JL, Furgiuele J, Wenz B. False positive immunometric assays be aware of their potential effect on immunometric assays caused by anti-immunoglobulin antibodies: a case report. Clin Chim Acta 147: 241, 1985. that utilize mouse monoclonal antibodies. 12) Cusick CF, Mistry K, Addison GM. Interference in a two-site Acknowledgement: Grateful acknowledgement is made to Miss Yoko immunoradiometric assay for thyrotropin in a child. Clin Chem Kuwahara for her excellent technical assistance. 31:348, 1985. 13) Thompson RJ, Jackson AP, Langlois N. Circulating antibodies to mouse monoclonal immunoglobulins in normal subjects - in References cidence, species specificity, and effects on a two-site assay for creatine kinase-MB isoenzyme. Clin Chem 32: 476, 1986. 14) Zweig MH, Csako G, Benson CC, Weintraub BD, Kahn BB. Hunter WM, Budd PS. Circulating antibodies to ovine and bovine Interference by anti-immunoglobulin G antibodies in immuno immunoglobulin in healthy subjects: a hazard for immunoassays. Lancet ii: 1136, 1980. radiometric assays of thyrotropin involving mouse monoclonal Prince AM, Brotman B, Jass D, Ikram H. Specificity of the direct antibodies. 15) Csako G, Weintraub BD, Zweig MH. The potency of immuno solid-phase radioimmunoassay for detection of hepatitis-B antigen. Lancet i: 1346, 1973. globulin G fragments for inhibition of interference caused by anti Howanitz PJ, Lamberson HV, Howanitz JH, Ennis KM. Circulating immunoglobulin antibodies in a monoclonal immunoradiometric human antibodies as a cause of falsely depressed ferritin values. assayClin for thyrotropin. Clin Chem 34: 1481, 1988. 16) Kahn BB, Weintraub BD, Csako G, Zweig MH. Factitious eleva Biochem 16: 341, 1983. Vladutiu AO, Sulewski JM, Pudlak KA, Stull CG. Heterophilic tion of thyrotropin in a new ultrasensitive assay: implications for the use of monoclonal antibodies in"sandwich" immunoassay. J antibodies interfering with radioimmunoassay. JAMA 19: 2489, Clin Endocrinol Metab 66: 526, 1988. 1982. Hedenborg G, Pettersson T, Carlstrom A. Heterophilic antibodies 17) Iitaka M, Ishii J, Ito K. Spurious serum TSH values due to anti causing falsely raised thyroid-stimulating hormone result. Lancet mouse IgG antibody of IgM class. Clin Chem 38: 311, 1992. 18) Zweig MH, Csako G, Spero M. Escape from blockade of interfering ii: 755, 1979. Gendrel D, Feinstein M-C, Grenier J, et al. Falsely elevated serum heterophile antibodies in a two-site immunoradiometric assay for thyrotropin (TSH) in newborn infants: transfer from mothersthyrotropin. to Clin Chem 34:2589, 1988. Skovsted L, Hansen JM. Falsely raised serum thyrotropin with infants of a factor interfering in the TSH radioimmunoassay.19) J Clin modified immunoradiometric assay. Lancet i: 359, 1988. Endocrinol Metab 52: 62, 1981. 20)hyper Harvey RD, McHardy KC, Trainer PJ, Reid I. Interference in Schaison G, Thomopoulos P, Moulias R, Feinstein MC. False thyrotropinemia induced by heterophilic antibodies against rabbit modified immunoradiometric assay for thyrotropin. Lancet i: 716, 1988. serum. J Clin Endocrinol Metab 53:200, 1981. Iitaka M, Tanikawa T, Sakatsume Y, Yanagisawa M, Hara Y, Ishii Czernichow P, Vandalem JL, Hennen G. Transient neonatal 21) hyper thyrotropinemia: a factitious syndrome due to the presence ofJ. Interference with thyrotropin receptor antibody determination by a spuriously occurring anti-bovine TSH antibody. Acta Endo heterophilic antibodies in the plasma of infants and their mothers. J Clin Endocrinol Metab 53: 387, 1981. crinol (Copenh) 112: 197, 1986.
988
Internal
Medicine
Vol. 31, No. 8 (August
1992)
ARTICLE
å¡
Incidence of Anti-Mouse IgG in Normal Subjects and Patients with Autoimmune Thyroid Disease Makoto Iitaka, Shinji Kitahama, Nobuhiko Fukasawa, Shiro Miura, Yoshito Kawakami, Shinichi Sakurai, Morifumi Yanagisawa and Jun Ishii
Serum antibodies to mouse IgG occasionally interfere with two-site immunometric assays in which mouse monoclonal antibodies are used. We examined the titers of antibodies to mouse IgG in serum samples from normal subjects and patients with autoimmune thyroid disease (AITD) using an enzyme-linked immunosorbent assay. Anti-mouse IgG antibodies were positive in 7/119 patients with Graves' disease (5.9%) and 3/60 patients with Hashimoto's thyroiditis (5.0%). One of the 15 patients with a thyroid neoplasm (6.7%) also had antibodies to mouse IgG, as did 5/60 healthy subjects (8.3%). These antibodies were either of the IgG or IgM class. There was no significant difference in the incidence of positive anti-mouse IgG antibody between normal subjects and patients with AITD. It is important to note this high incidence of antibodies to mouse IgG due to the potential of interference with immunometric assays employing mouse monoclonal (Internal Medicine 31: 984-988, 1992) antibodies. Key words: heterophilic antibody, two-site immunometric assay
Introduction
for anti-mouse IgG antibodies. We found that a con siderable number of patients with AITD had antibodies to mouse IgG. However, there was no significant difference in the incidence of antibodies to mouse IgG between the AITD patients and the control subjects.
Heterophilic antibodies are known to interfere with numerous of important immunoassays, such as those for or-fetoprotein (1), hepatitis B surface antigen (2), ferritin (3), hCG (4), and TSH (5-9). A 7% incidence of antibodies Materials and Methods against ovine or bovine immunoglobulin has been reported in healthy individuals (1). Antibodies to the immuno Su bjects globulins of various animal species, such as rabbit (3, 5-9), One-hundred and nineteen patients with Graves' disease guinea pig (2), and goat (4) have also been found. The recent (21 men and 98 women) and 60 patients with Hashimoto's development of two-site immunometric assays using mouse thyroiditis (8 men and 52 women) were examined. As monoclonal antibodies has enabled higher specificities and controls, 15 patients with thyroid tumors (10 with benign sensitivities. However, serum antibodies to mouse IgG may adenoma and 5 with cancer, 2 men and 13 women), and interfere with many of the assay kits employing mouse 60 apparently healthy subjects (18 men and 42 women) who monoclonal antibodies (10-16). To prevent such inter did not have antibodies to thyroglobulin or thyroid ference, nonimmune mouse serum or IgG are commonly microsomes were also studied. The serum was separated included in these kits. However, we previously encountered immediately after the blood withdrawal and stored at a patient with Graves' disease whose serum interferred with -20°C. the TSH assay kit that included mouse IgG (17). This Measurement of anti-mouse IgG antibodies prompted us to examine the incidence of serum antibodies The titers of antibodies to mouse IgG were measured to mouse IgG in patients with autoimmune thyroid disease using an enzyme-linked immunosorbent assay (ELISA). One (AITD). As "nonimmune" controls, serum from healthy From the Fourth Department Internal Medicine, Saitama individuals and patients withof thyroid tumors was also testedMedical School, Saitama Received for publication December 4, 1991; Accepted for publication May 7, 1992 Reprint requests should be addressed to Dr. Makoto Iitaka, the Fourth Department of Internal Medicine, Saitama Medical School, 38 Mo Moroyama, Iruma-gun, Saitama 350-04, Japan 984
Internal
Medicine
Vol. 31,
No. 8 (August
1992)
Incidence
of Anti-Mouse IgG Antibody
milligram of purified mouse IgG (Sigma Chemical Co. , St. Louis, MO) was labelled with 60 ^g of N-hydroxysuccini midobiotin (Pierce, Rock ford, IL). Ninety-six well plates (Immunoplate, Nunc, Denmark) were coated with 10 mg/1 of purified mouse IgG in 0. 1 mol/1 carbonate-bicarbonate buffer, pH 8.3. Plates were then incubated with 1 g/1 bovine serum albumin (Sigma) in phosphate-buffered saline (PBS) to reduce nonspecific binding. Serum samples (50/ul) were incubated in the plates for 1 hour at room temperature. The plates were then washed 3 times with PBS containing 0. 1 % Tween 20 (washing buffer), biotinylated mouse IgG was added, and the plates were incubated at room temperature for 1 hour. After washing 3 times with washing buffer, the plates were incubated with avidin-conjugated horseradish peroxidase (Sigma) at room temperature for 1 hour. The plates were then washed an additional 3 times with washing buffer, and 0.4 g/1 of o-phenylenediamine (Wako Chemical Co., Osaka, Japan) in 0.1 mol/1 citrate buffer, pH 4.8, con taining 0.003% hydrogenperoxide was added. After 30 minutes of incubation at room temperature, 50#1 of 6 mol/1 H2SO4 was added to stop the reaction. The optical density (O.D.) at 492 nm was then measured using an ELISA reader. Affinity-purified goat (IgG) antibodies to mouse IgG (Cappel, West Chester, PA) were diluted to 1:10,000 and used as a standard. Titers for the anti-mouse IgG antibodies were expressed as the ratio of the O.D. of the serum sample to that of the standard (ELISA index). To examine the
solution containing 250 mg/1 of 3,3'diaminobenzidine (Wako) and 0.01 % hydrogenperoxide in 0.1 mol/1 Tris HC1 buffer, pH 7.4. Immunoglobulins bound to biotinylated mouse IgG were detected as a brown precipitation line. Laboratory measurements Serum thyroid hormone and TSH levels were measured using commercially available kits (Amerlex-M, Amersham International Pic, Buckinghamshire, England; and RIA gnost hTSH, Behringwerke, Marburg, FRG, respectively). The normal ranges of serum free thyroxine (FT4) and TSH were 9.7-24.5 pmol/1 and 0.3-3.5 mU/1, respectively. Antibodies to thyroglobulin (TGPA) and thyroid micro somes (MCPA) were detected using the kits obtained from Fuji Rebio (Tokyo, Japan). TSH receptor antibodies (TRAb) were determined by a radioreceptor assay (S.R.S. Ltd, Cardiff, UK). Statistical analysis The Smirnov-Grubbs test was used to exclude extremely deviated data in the determination of the normal range. The chi-squared test was used to compare the incidence of serum antibodies to mouse IgG in the healthy subjects and in the patients with thyroid disease. Results
Specificity and accuracy of the ELISA for anti-mouse IgG an tibodies To evaluate the specificity of this assay, bovine, rabbit, and mouse IgG were added to serum samples from four specificity of the assays, varying amounts of rabbit, bovine, subjects positive for antibodies to mouse IgG. As shown Immunoelectrophoresis and enzymatic detection of anti in Table 1, up to 500fig of bovine and rabbit IgG and no and mouse IgG (Sigma) were added to serum samples mouse IgG antibodies effect on the assay. On the other hand, 50^g of mouse IgG obtained from the subjects for anti-mouse IgG Immunoelectrophoresis waspositive performed as previously significantly reduced the readings. The intra-assay co antibodies(17, and 18) the using ELISAgoat wasanti-human repeated. whole serum, anti described efficient of variation was 1.6-ll.8% and the interassay human IgG or anti-human IgM (Cappel). Biotinylated variation was 3.7-24.1%. Of the 60 serum samples ob mouse IgG was then applied to the plates. After overnight tained from healthy subjects, 1 gave an extraordinarily high incubation, the plates were washed extensively with washing ELISA index. This was excluded by the Smirnov-Grubbs buffer and incubated with avidin-conjugated horseradish test when determining the normal range. The mean + 2 SD peroxidase (Sigma) for 2 hours at room temperature. After of the remaining 59 samples was considered as the upper further washing, the plates were incubated with the substrate limitonofthe normal thisAnti-Mouse study. Samples Table 1. Inhibitory Effect of Various IgG Assayinfor IgG with an ELISA index Antibodies A dd ition of ( ｫg) S ubject
R abbit IgG 100 250 500
100
B ovin e IgG 250 500
M o use IgG 50
C ontro l
0 .583
0 .617
0 .479
0.500
0.625
0.502
0.0 18
0 .521
0 .354 0 .646
0 .429 0 .667
0 .354 0 .546
0.542 0.646
0.4 17 0.583
0.407 0.576
0.023 0.024
0 .4 72 0 .597
0 .621
0 .633
0 .52 1
0.59 8
0.5 12
0.503
0.02 8
0 .576
The ELISA index for anti-mouse IgG antibody was determined in the presence or absence (control) of IgG from various animal species. Internal
Medicine
Vol. 31,
No. 8 (August
1992)
985
Iitaka Table 2.
Incidence
et al
of Anti-Mouse IgG Antibodies
T o ta l
E L IS A in d e x
N u m b er o f
In cid e n c e o f
S u b jects
n um ber
(m e a n ｱ S D )
p o sitiv e p a tien ts
p o sitiv ity (% )
G D
1 19
0 .2 16 + 0 .13 7
7
5 .9
H T T N
60 15 60
0 .2 10 + 0 .12 1 0 .2 2 9 + 0 .14 0 0 .17 8 + 0 .1 16
3 1 5
5 .0 6 .7 8 .3
GD: Graves' disease, HT: Hashimoto's thyroiditis, TN: thyroid neoplasm, C: healthy control. *An ELISA index above the mean + 2 SD for 59 healthy subjects was defined as positive for anti-mouse IgG antibody, [one extremely high ELISA index value determined in a healthy subject was excluded by the Smirnov-Grubbs test.] Table 3. Clinical Profile of the Patients with Autoimmune Thyroid Disease Who were Positive for Anti-Mouse IgG Antibody P atient D x
E L ISA ind ex F T 4
TSH
TR A b
N o. 1 G D 2 GD 3 HT
0 .535 1.405 0 .660
15.3 8.5 14.2
< 0 .1 < 0.1 0.26
53 .8 55 .8 4 .5
4 5 6 7
0 .428 0 .432 0 .447 0 .590 0 .9 11 0 .569 0 .4 16
6.2 < 0 .1 13.1 1.2 7.3 0.85 20.7 2 .9 15.7 1.3 21.6 2.6 14.2 2 .7
30 .3 - 1.6 - 67.6 6 .5 5 .0 5 .7 0 .0
9 10
G G G G H G H
D D D D T ,R A D T
A TA
AM A
Ig class#
6 ,400 25 ,600 6 ,400
IgG IgG IgM
1,600 400 1,600 1,600 6 ,400 6 ,400 102 ,400 102 ,400 2 5,600 400 6 ,400 6 ,400
ND ND ND IgG IgG ND ND
100 400
Dx: diagnosis, GD, Graves' disease, HT: Hashimoto's thyroiditis, TRAb: TSH receptor antibody, ATA: anti-thyroglobulin antibody, AMA: anti-microsomal antibody, ND: notclass determined. #Immunoglobulin of anti-mouse IgG antibody. *TSH was measured with the RIAgnost TSH kit. Normal range: FT4: 9.7-24.5 pmol/1, TSH: 0.3-3.5 mU/1 TRAb: -10 to +10% above this value (>0.409) were defined as positive for anti-mouse IgG antibodies. A nti-mouse IgG antibodies in patients with thyroid disease and in healthy subjects Seven out of the 1 19 patients with Graves' disease (5.9%) had anti-mouse IgG antibodies. Three of the 60 patients with Hashimoto's thyroiditis (5.0%) and 1 of the 15 patients with thyroid neoplasms (6.7%) were also positive for these anti bodies. It is of interest that 5 of the 60 healthy subjects (8.3%) were also positive for anti-mouse IgG antibodies. As shown in Table 2, there was no significant difference in the mean ELISA index and the incidence of these anti bodies between the patients and the normal subjects. Clinical details of the patients with AITD who were positive for anti-mouse IgG antibodies are shown in Table 3. There was no correlation between the ELISA index for the anti-mouse IgG antibodies and thyroid autoantibodies or serum thyroid hormone levels in the patients with AITD. Furthermore, there was no prominent discrepancy between serum TSH and free T4 (FT4) levels, suggesting that the anti-mouse 986
antibodies in these patients did not interfere with the TSH assay used in this study. It is of interest that Graves' disease patient No. 6 had an abnormally negative value for TRAb due to the presence of serum anti-bovine TSH antibodies. Examination of longitudinally collected serum samples from two patients with Graves' disease showed that the anti mouse IgG antibodies were positive throughout the course irrespective of the serum levels of thyroid hormone or autoantibodies, although the ELISA index decreased with time in both cases (Table 4). Immunoglobulin class of anti-mouse IgG antibodies Immunoelectrophoresis and subsequent enzymatic de tection showed that antibodies to mouse IgG were either ofthe IgG or IgM class (Fig. 1). As shown in Table 3, 4 patients with AITD had anti-mouse antibodies of the IgG class. One patient with Hashimoto's thyroiditis had anti mouse IgG antibodies of the IgM class.
Internal
Medicine
Vol. 31, No. 8 (August
1992)
Incidence
of Anti-Mouse IgG Antibody
Table 4. Longitudinal Changes of the Titers of Anti-Mouse IgG suggesting that the nonimmune mouse serum or IgG in the Antibodies, the Titers of Thyroid Autoantibodies, and the kits used in this study effectively inhibited interference by Serum Thyroid Hormone Levels in the Patients with Graves' the serum antibodies. Previous studies (18-20) have shown Disease that antibodies to mouse IgG interfere with some, but not P a tie n t D a te (Y / M / D ) E L IS A in d e x FT 4 TR A b A TA A M A all, immunometric assays. This may be due to differences in the immunoglobulin class or subclass of the mouse M M 851225 0 .9 3 3 5 .3 N A 100 6 ,4 00 860605 0 .8 8 9 1 7 .9 N A 1 ,6 0 0 monoclonal antibodies in the kits or to differences in the 870525 0 .8 3 7 1 2 0 .3 4 1 .9 1 ,6 0 0 absorbing capacity of the nonimmune mouse serum or IgG 880609 0 .7 2 6 1 5 .7 3 9 .8 6 ,4 0 0 in the kits. Indeed, antibodies to mouse IgG found in our 890729 0 .6 9 3 1 4 .5 3 1 .2 1 ,6 0 0 previously reported patient did not interfere with the TSH 900602 0 .6 0 3 17 .6 1 8 .7 1 ,6 0 0 assay kit used in this study (17). These observations indicate SS 840605 0 .7 2 0 1 3 .O N A 1 0 2 ,4 0 0 that anti-mouse IgG antibodies detected in this study may 8 5 0 2 15 0 .6 7 7 1 7 .6 N A 2 5 ,6 0 0 have the potential to interfere with some other 860225 0 .6 1 3 2 9 .2 N A 400 2 5 ,6 0 0 Antibodies to mouse IgG were consistently detected in 870525 0 .4 3 1 1 2 .9 13 .2 10 0 4 0 9 ,6 0 0 immunometric assays. 880527 0 .5 1 6 1 5 .4 3 .1 10 0 6 ,4 0 0 some Graves' patients throughout the course of their 890203 0 .4 0 6 1 5 .6 6 .7 10 0 1 0 2 ,4 0 0 disease, regardless of their thyroid function or thyroid 900323 0 .4 4 7 19 .3 4 .2 2 5 ,6 0 0 autoantibody titers. There was thus no relation between TRAb : TSH receptor antibody, ATA: anti-thyroglobulin antibody, AMA: for anti-mouse IgG antibodies and thyroid auto positivity anti-thyroid microsomal antibody, NA: not available. antibodies. Furthermore, antibodies against mouse IgG were detected almost equally in the healthy subjects and the patients with thyroid neoplasms as in the patients with AITD.These observations suggest that the production of antibodies to mouse IgG is not related to thyroid auto immunity. It is of interest that one patient with Graves' disease (patient No. 6 in Table 3) had antibodies to both mouse IgG and bovine TSH. Antibodies to bovine TSH interfere with the TSH receptor antibody assay and cause extremely negative TRAb values (21). It is interesting to speculate about thistopatient develop such The incidence ofwhether antibodies sheep is orprone bovinetoimmuno heterophilic antibodies. globulin has been reported to be 7^o in a population of blood donors (1). Hedenborg et al (5) have suggested that at least 5% of all sera tested were positive for antibodies to rabbit immunoglobulin. The incidence of previously reported heterophilic antibodies thus appears to be quite Fig. 1. Immunoglobulin class of antibodies to mouse IgG. Immuno similar to that for anti-mouse IgG antibodies shown in this globulins were precipitated with antibodies to human whole serum. Immunoglobulins bound to biotin-labelled mouse IgG could be seen after study (5.0-8.3%). Our results are also compatible with the the enzymatic reaction with avidin-peroxidase and the substrate. study Px andby Thompson et al (13), who reported that about 9% P2 (patients No. 1 and 2 in Table 3) show the presence of antibodies of of blood the IgG class directed against mouse IgG. P3 (patient No. 3 in Table 3) donors had spuriously high CK-MB isoenzyme indicates the presence of an antibody of the IgM class. N is a normal values which were greatly depressed when re-assayed in the control. presence of nonimmune mouse serum. It may seem sur prising that the incidence of heterophilic antibodies is fairly Discussion high, although the incidence of interference with an assay We theantiserum immunoglobulin classto ofbe theonly anti thatalso usesexamined guinea pig was reported We previously encountered a patient with Graves' disease 0.52% (3). mouse IgG antibodies. To our knowledge, this is the first whose serum interfered with the TSH assay in which mouse report that demonstrated the presence of IgG and IgM IgG was included (17). Similar to our case, other patients classes of anti-mouse IgG antibodies. No significant who escaped from such blockade have also been reported difference was found in thyroid function tests or thyroid (18-20). In the present study, we showed that 5-6% of autoantibody titers between patients with IgG class anti bodies and those with IgM class antibodies. It remains to patients with AITD had antibodies to mouse IgG. However, be clarified whether or not there is class switching of these the patients with anti-mouse IgG antibodies had serum TSH antibodies. and thyroid hormone levels compatible with clinical status , Internal
Medicine
Vol. 31, No. 8 (August
1992)
987
Iitaka
et al
In the present study, we demonstrated that antibodies 9) Howanitz PJ, Howanitz JH, Lamberson HV, Ennis KM. Incidence and mechanism of spurious increase in serum thyrotropin. Gin Chem against mouse IgG were present in 6-8% of both patients 28: 427, 1982. with thyroid disease and healthy subjects. Although the 10) Bock JL, Furgiuele J, Segen JC. Choriogonadotropin measured with the Tandem-E immunoenzymetric assay system. Clin Chem 31:441 , precise mechanism of the production of these anti-mouse 1985. IgG antibodies remains to be clarified, it is important to 1 1) Bock JL, Furgiuele J, Wenz B. False positive immunometric assays be aware of their potential effect on immunometric assays caused by anti-immunoglobulin antibodies: a case report. Clin Chim Acta 147: 241, 1985. that utilize mouse monoclonal antibodies. 12) Cusick CF, Mistry K, Addison GM. Interference in a two-site Acknowledgement: Grateful acknowledgement is made to Miss Yoko immunoradiometric assay for thyrotropin in a child. Clin Chem Kuwahara for her excellent technical assistance. 31:348, 1985. 13) Thompson RJ, Jackson AP, Langlois N. Circulating antibodies to mouse monoclonal immunoglobulins in normal subjects - in References cidence, species specificity, and effects on a two-site assay for creatine kinase-MB isoenzyme. Clin Chem 32: 476, 1986. 14) Zweig MH, Csako G, Benson CC, Weintraub BD, Kahn BB. Hunter WM, Budd PS. Circulating antibodies to ovine and bovine Interference by anti-immunoglobulin G antibodies in immuno immunoglobulin in healthy subjects: a hazard for immunoassays. Lancet ii: 1136, 1980. radiometric assays of thyrotropin involving mouse monoclonal Prince AM, Brotman B, Jass D, Ikram H. Specificity of the direct antibodies. 15) Csako G, Weintraub BD, Zweig MH. The potency of immuno solid-phase radioimmunoassay for detection of hepatitis-B antigen. Lancet i: 1346, 1973. globulin G fragments for inhibition of interference caused by anti Howanitz PJ, Lamberson HV, Howanitz JH, Ennis KM. Circulating immunoglobulin antibodies in a monoclonal immunoradiometric human antibodies as a cause of falsely depressed ferritin values. assayClin for thyrotropin. Clin Chem 34: 1481, 1988. 16) Kahn BB, Weintraub BD, Csako G, Zweig MH. Factitious eleva Biochem 16: 341, 1983. Vladutiu AO, Sulewski JM, Pudlak KA, Stull CG. Heterophilic tion of thyrotropin in a new ultrasensitive assay: implications for the use of monoclonal antibodies in"sandwich" immunoassay. J antibodies interfering with radioimmunoassay. JAMA 19: 2489, Clin Endocrinol Metab 66: 526, 1988. 1982. Hedenborg G, Pettersson T, Carlstrom A. Heterophilic antibodies 17) Iitaka M, Ishii J, Ito K. Spurious serum TSH values due to anti causing falsely raised thyroid-stimulating hormone result. Lancet mouse IgG antibody of IgM class. Clin Chem 38: 311, 1992. 18) Zweig MH, Csako G, Spero M. Escape from blockade of interfering ii: 755, 1979. Gendrel D, Feinstein M-C, Grenier J, et al. Falsely elevated serum heterophile antibodies in a two-site immunoradiometric assay for thyrotropin (TSH) in newborn infants: transfer from mothersthyrotropin. to Clin Chem 34:2589, 1988. Skovsted L, Hansen JM. Falsely raised serum thyrotropin with infants of a factor interfering in the TSH radioimmunoassay.19) J Clin modified immunoradiometric assay. Lancet i: 359, 1988. Endocrinol Metab 52: 62, 1981. 20)hyper Harvey RD, McHardy KC, Trainer PJ, Reid I. Interference in Schaison G, Thomopoulos P, Moulias R, Feinstein MC. False thyrotropinemia induced by heterophilic antibodies against rabbit modified immunoradiometric assay for thyrotropin. Lancet i: 716, 1988. serum. J Clin Endocrinol Metab 53:200, 1981. Iitaka M, Tanikawa T, Sakatsume Y, Yanagisawa M, Hara Y, Ishii Czernichow P, Vandalem JL, Hennen G. Transient neonatal 21) hyper thyrotropinemia: a factitious syndrome due to the presence ofJ. Interference with thyrotropin receptor antibody determination by a spuriously occurring anti-bovine TSH antibody. Acta Endo heterophilic antibodies in the plasma of infants and their mothers. J Clin Endocrinol Metab 53: 387, 1981. crinol (Copenh) 112: 197, 1986.
988
Internal
Medicine
Vol. 31, No. 8 (August
1992)
